By Carlos Galamba, vice president of IVD intelligence and innovation, RQM+
Notified body feedback reveals that the biggest gaps in IVDR submissions are being found in clinical evidence, particularly around clinical performance data. Under the EU’s In Vitro Diagnostic Medical Devices Regulation 2017/746 (IVDR), clinical evidence must demonstrate a continuous process of performance evaluation and must support the intended purpose of the device. The first wave of IVDR assessments shows that in many cases, the intended purpose is poorly defined and is not aligned with the clinical performance data that is being presented. As a result, notified bodies are finding a lack of support for the claims and indications made and, where applicable, the clinical benefit(s) specified in the intended purpose.
Manufacturers may be underestimating how much effort is truly involved in the performance evaluation process. The IVDR greatly increases the number of IVD devices subject to regulatory oversight and expands the overall scope of compliance. Since there is no grandfathering, legacy devices also need to meet performance evaluation requirements. With only seven IVDR-designated notified bodies so far, and many at capacity already, devices that show significant gaps in clinical evidence risk being refused, particularly if those gaps cannot be closed in the timeframes stipulated for the notified body review. This could ultimately threaten a manufacturer’s IVDR transition and disrupt market access for their product portfolio.
To avoid this scenario, defining a realistic clinical evidence strategy for IVDR compliance and revisiting existing data is a must. If a company sells diagnostic products that are 510(K) cleared in the U.S., some of the required evidence may already be available. Manufacturers can leverage existing performance evaluation data to facilitate the IVDR process. As well as providing examples of overlap between FDA and IVDR requirements, this article outlines lessons learned on IVDR performance evaluation with a focus on the greatest challenges manufacturers are facing, such as clinical performance.
The 3 Clinical Evidence Pillars
Performance evaluation is a process used to generate clinical evidence and is carried out according to a performance evaluation plan. This is where the depth of evaluation and the level of clinical evidence required are planned to consider several factors, including (non-exhaustive):
- device risk
- novelty and degree of innovation
- intended purpose and the specific claims being made
- the state-of-the-art
- intended user
- target population
The process involves the assessment and analysis of data to establish or verify scientific validity, analytical performance, and, where applicable, clinical performance. The sum total of these three pillars is clinical evidence, which is summarized in a performance evaluation report (PER). The aim of the PER is twofold: first, to demonstrate overall conformity with the General Safety and Performance Requirements and, second, to scientifically demonstrate, by reference to state-of-the-art in medicine, that the in vitro diagnostic device (IVD) is safe and that the intended clinical benefit is achieved.
Scientific validity refers to the association between an analyte and a clinical condition or physiological state. It can be demonstrated using peer-reviewed scientific literature, clinical guidelines from professional associations, proof of concept studies, clinical performance studies, or consensus expert opinions, for example. It is advisable to demonstrate that both favorable and unfavorable publications have been considered. Manufacturers can also draw on relevant information on the scientific validity of devices measuring the same analyte or biomarker.
Analytical studies are always required for analytical performance. Typically, CLSI (Clinical and Laboratory Standards Institute) guidelines are used to demonstrate analytical performance. Studies must address all the applicable parameters in point (a) of Section 9.1 of Annex I of the IVDR. Any parameters deemed not applicable should be justified in performance evaluation documents. This includes, for example, analytical sensitivity and specificity, accuracy, limit of detection, and linearity.
Clinical performance tends to focus on the patient, the clinical condition, or physiological state. Data can be sourced from clinical performance studies, scientific peer reviewed literature, and/or published experience from routine diagnostic testing. The selected approach may vary, depending on novelty, device classification, and risk. The IVDR does not define what “published experience gained by routine diagnostic testing is,” but this is understood as being data from proficiency testing schemes, external quality assessment (EQA), and ring-trials (inter-laboratory comparison), among others.
Notified Body Expectations And Frequent Feedback
Clinical Evidence Approach
The PER needs to include all elements outlined in Part A of Annex XIII 1.3.2, and one of the issues most frequently found lies in the description of the approach taken to gather clinical evidence, in particular the approach for clinical performance (CP). This is often poorly described and not aligned with IVDR terminology, missing a link to the CP sources required by the IVDR in part A of Annex XIII 1.2.3 (i.e., clinical performance studies, scientific peer-reviewed literature, published experience from routine diagnostic testing, and other sources of clinical performance data). Manufacturers also miss justifications for not conducting CP studies and call their legacy/old studies IVDR clinical performance studies, which is incorrect, as we will discuss later in this article.
Intended Purpose Statements, Claim, And Limitations
Poorly defined and/or too broad intended purposes and claims are also difficult to prove, exposing the PER to unwanted gaps. Intended purposes that link to several clinical conditions or have few limitations (e.g., all types of cancer) are particularly challenging, as each individual claim will need to be traceable in the PER and supported by relevant clinical data. However, one should note that many biomarkers are not specific to a particular clinical condition and establishing clinical performance data on a condition basis may not be practical.
Let’s look at C-Reactive Protein (CRP), which is a general marker of infection and inflammation and is elevated in a variety of conditions from autoimmune diseases to some forms of arthritis and inflammatory bowel disease. For a general biomarker like CRP, it would not be reasonable to conduct separate clinical studies for each associated condition and/or clinical setting. Instead, scientific validity would be the right place to demonstrate the relevance of CRP in all referenced settings, with performance focused on the ability for the assay to accurately detect CRP, through analytical performance and method comparison studies with a predicate device, for example. However, it all comes down to the intended purpose statement and the claims made. Let’s say that same CRP marker is claiming a prognostic value around the clinical course of a patient with sepsis. In this scenario, clinical performance data focused on prognosis would most likely be required.
The intended purpose should also include specimen claims that are backed up by the actual data obtained in the relevant studies. For example, claims of nasal vs. nasopharyngeal sample types in SARS-CoV-2 antigen assays should be supported by actual studies with those specimen types.
Finally, notified bodies are also scrutinizing claims where there is no limitation of the target population. For example, if clinical performance data was obtained with adult samples only, a limitation to this demographic would be expected in the instructions for use.
Was Clinical Performance Data Obtained With The Final User?
The level of detail also depends on who is using the product and the use environment. IVDs may not always be used in a laboratory by highly trained personnel. They can also be used for “near-patient testing”, where tests are performed outside the laboratory, such as in an ambulance by a paramedic or at the patient’s bedside by a nurse. Since the PER must demonstrate that performance evaluation was carried out with the intended user in mind, clinical performance data for a near-patient test should be obtained with the range of potential users. These should be clearly defined in the technical documentation. For all IVDs, but particularly for near-patient tests and self-tests, it is important to show that usability factors have been considered, such as human factors, adequacy of the instructions for use, and specific training requirements.
What To Expect From Notified Bodies
Manufacturers should be aware that the higher the risk, the higher notified body expectations will be for clinical evidence. Looking at clinical performance, for example, manufacturers are typically expected to conduct clinical performance studies, unless they can confidently rely on other sources, such as peer-reviewed literature or published experience gained by routine diagnostic testing. These other sources may be sufficient for lower risk and some legacy devices, but they would not stand up to scrutiny for a high-risk novel device.
Nevertheless, irrespective of classification, the decision to not conduct a clinical performance study needs to be duly justified. Ultimately, the PER should demonstrate that even without a clinical performance study, the IVD still achieves the intended clinical benefit(s) and safety when used as intended, based on the CP sources selected by the manufacturer.
The recently published MDCG 2022-2 guidance on the general principles of clinical evidence for IVDs further clarifies that old clinical studies conducted under the IVDD do not replace an IVDR clinical performance study since they do not meet the requirements of Annex XIII 2.3. These studies can still be used but they would need to be presented as “other sources of clinical performance data” per Annex XIII 1.2.3.
This does not mean that manufacturers are not allowed to use their existing clinical studies predating IVDR, and it certainly does not mean that every legacy high-risk device needs to be supported by an Annex XIII 2.3. Clinical Performance Study Plan and Report. What is important is that manufacturers assess the quality and completeness of their existing data to identify any potential gaps and define a strategy to close those gaps.
EU & U.S. Overlap
A thorough review of existing data may produce some of the performance evaluation data needed for IVDR compliance. In fact, products that are FDA 510(K) cleared may have a head start, since a significant proportion of supporting testing and performance data may have already been sourced. After reviewing existing data, manufacturers can then source the study reports used to support the FDA submission to build a complete picture of the data and get a clearer idea of gaps that need filling. Manufacturers can use this process to consolidate regulatory information and systems, which can only be beneficial for future compliance activities and maintenance.
One area where requirements are fairly consistent is analytical performance (AP). Under both regulatory frameworks, the parameters that must be addressed are largely similar, with both the U.S. and the EU relying on state-of-the-art CLSI guidelines for AP studies.
For clinical performance, notified bodies will typically expect studies with samples that are representative of Europeans, unless due justification is given for using specimens from other geographies. Clinical performance studies under the IVDR also need to follow well-established standards, such as ISO 20916 on clinical performance studies using specimens from human subjects, and must adhere to the requirements outlined in Annex XIII section 2, including the contents outlined for the Clinical Performance Study Plan and Report (CPSP and CPSR).
Overall, one of the major changes IVDR brings is how prescriptive performance evaluation requirements have become and although much of FDA’s 510(K) data can be used, manufacturers still need to present performance evaluation documentation that meets all requirements of Annex XIII. The efforts involved in generating this new body of clinical evidence and all required deliverables cannot be underestimated.
The new classification rules of the IVDR mean that an estimated 84% of IVDs now fall under notified body scrutiny, compared to only 7% under the directive.1 At the same time, notified body expectations are higher, compared to the outgoing EU In Vitro Diagnostic Directive. A well-planned strategy will therefore minimize the need for remediation following initial submission and help to avoid the very real danger of product withdrawal. Though some products can benefit from a longer transition period following the date of application in May 2022, it is highly recommended that you do not delay your compliance efforts, as you are likely to need all the time available. Gathering clinical evidence is one of the most challenging areas of IVDR compliance, so prioritizing this is likely to be beneficial.
For further guidance on clinical evidence requirements under the IVDR, including a comparison between EU and U.S. requirements, download the free RQM+ white paper on the topic here.
About The Author:
Carlos Galamba is vice president of IVD intelligence and innovation at RQM+. He provides thought leadership and helps to drive technical quality for in vitro diagnostics. Previously, he spent seven years at BSI, where he was responsible for managing, coaching, and developing a global team of IVD technical experts. He was BSI’s first in-house clinician for IVDs, leading the implementation of the BSI clinical oversight process and making hundreds of CE marking recommendations for IVDs.
1The impact of the new European IVD classification rules on the notified body involvement; a study on the IVDs registered in the Netherlands RIVM Letter report 2018-0082 A. van Drongelen et al.