ISO 15189 is a sufficient instrument to guarantee high-quality manufacture of laboratory developed tests for in-house-use conform requirements of the European In-Vitro-Diagnostics Regulation

1.1 Timelines and impact

Entry into force of the IVDR [4] was planned for May 26 2022. The European Federation of Clinical Chemistry and Laboratory Medicine (EFLM) commented on uncertainty of continued availability of commercial legacy devices and on the impact of implementing the IVDR for LDTs [7]. The European Community legislator published an amending act [8] extending the timelines (Figure 1). The introduction of the REACH regulation [9] may force redesign of devices with respect to accepted reagents and additives, potentially affecting the performance of the devices. This will also add to the workload. The MDCG follows up the roll-out of the elements of the new regulatory framework and their capacity to manage the transition of IVDD legacy devices to avoid market disruptions and a collapse of diagnostic services [10]. Despite a will across the sector to seek workable solutions, the obstacles remain formidable, and the potential solutions so far proposed remain more a matter of aspirations than of clear pathways [11].

Figure 1:

Protracted timeline for IVDR-compliant implementation of in-house laboratory-developed tests (LDTs). This is a non-authoritative abridged interpretation of the amending EU regulation 2022/112 [8]. As of the date of submission of our paper, the revised timelines are as indicated in the Figure. Commercial devices put into use can be used beyond the times indicated in the Figure, as long as they are within the applicable expiration period. The extended timeline for commercial devices results in the extension for the IVDR 5 (5) d requirement on the absence of marketed equivalent devices. With respect to LDTs: IVDR 5 (5) g applies to class D devices, but member states can extend this at will to other classes.

In typical clinical laboratories, over 95% of routine laboratory results are produced using IVDD and/or IVDR-conform devices, while the remainder is produced by LDTs for which there is no commercial alternative, or commercial devices only fulfill the laboratories requirements after extension of the intended use or by modification of the design, or commercial devices cannot be run on available platforms [12, 13].

For large volume commodity tests such as routine chemistry and hematology tests, the LDTs will correspond to a small fraction of the bulk revenue generated by the laboratory. Other specialized tests may have a market share that is too small to interest economic actors. These tests will entirely depend on LDTs. In rare or orphan diseases, the majority of tests make use of “for research use only” (RUO) components. Even when these tests are often concentrated in a few specialized laboratories, relatively small volumes and costly development might become prohibitively expensive. By the same token, we expect many IVDD legacy commercial IVDs with limited economic interest to disappear from the market, forcing the laboratories to develop LDT alternatives. Absent the numbers involved, the dates set forward in Figure 1 dictates the timeframe for their discontinuation. LDTs add to the workload for the HIs and CAs.

In summary, the laboratory has to bring about 50% of its test portfolio responsible for less than 5% of reported results into IVDR compliance [12]. Per the timelines summarized in Figure 1, this implies demonstrating absence of equivalent commercial devices that meet the target group’s needs at the appropriate level of performance (IVDR 5 (5) d). The HI has to validate the LDT according to IVDR ANNEX I, and to fulfill extra documentary requirements (IVDR5 (5)). For the highest risk class (IVDR ANNEX VIII), the HI has to draw up a document sufficiently detailed for its CA to be able to understand the manufacturing process and to ascertain whether the general requirements of IVDR ANNEX I are met. The devices cannot be manufactured on an industrial scale, and cannot be transferred to other legal entities (IVDR 5 (5) a). For all other LDTs the HI provides a justification for their manufacture upon request by the CA. The CA can inspect the manufacturing activities at the HI.

3.1 Preliminaries

3.1.2 Absence of device with the appropriate level of performance on the market

LDTs address a clinical or operational need not met by available devices on the market. In Figure 2 we propose an applicable decision tree.

Figure 2:

Weighing the LDT with respect to the IVDR 5 (5) d non-equivalence requirement. The flowchart summarizes decisions to be made when deciding about whether an application has to be treated as an LDT. The diamond in dashed lines corresponds to MDCG 2022-6 for legacy devices [27]. In the absence of equivalent guidance, we apply it here also to LDTs. Decisions ending in the “not-significant” path have to be documented internally. Although the obligation for documentation of absence of an equivalent commercial device is postponed for existing LDTs till 2028, this test precedes the diamond in dashed lines, because in the event of a modification of an IVDD-registered test after May 26 2022 the latter date of entering into force of the IVDR still applies to new LDTs.

The proposed scheme starts with a market exploration. With the deferred timelines mentioned in Section 1.1, Figure 1, the laboratory only has to demonstrate the absence of an equivalent device on the market from May 26, 2028. Meanwhile IVDs put onto the market under the old IVD directive are phased out. In the absence of a marketed device, or marketed devices not fulfilling the appropriate level of performance, we can resort to the LDT. When we decide to modify a marketed device, we have to justify the modification and its use (IVDR 5 (5) d-e). This requirement came into force from May 26 2022, although the documentation will only to be available from May 26 2024 [8].

Existing devices are weighted at our performance criteria. These include intended use, safety considerations, clinical and analytical performance and operational requirements. Operational requirements can be, but are not limited to following examples: turn-around time required by the patient care programs serviced, availability in the laboratory of technical platforms, and guaranteed supply.

If a change to a marketed device is required, then we have to evaluate whether such change is critical. A critical change will turn it into an LDT, requiring proper validation. By contrast a non-significant change does not have to be handled as a relevant deviation of the original device. We adopt here guidance on the significance of changes with respect to transitional provisions for commercial legacy devices [27]. Table 3 summarizes elements to consider in the risk-analysis to decide on the criticality of changes to devices. When evaluated as critical these elements will then have to be addressed while validating the LDT for use. Contrary to medical devices, diagnostic devices are seldomly used as a stand-alone element in diagnosis and or treatment selection. We propose to use Table 3 not as a prescriptive tool, but as a checklist allowing for educated choices. Under Section 4 we discuss examples of evaluation of criticality and commensurate actions. The analysis is independent from the risk classification according to ANNEX VIII. The role of tests and their position in clinical care pathways and the risk for and consequences of misdiagnosis or missed diagnosis has to be taken into consideration when evaluating criticality of changes.

Table 3:

Evaluating criticality of changes in existing methods.

A. Extension or change to the intended purpose – New user or patient population – Change of clinical use (anatomical site/sample matrix) B. Change in design or performance specification – Requires further clinical or usability data – New risks/existing risks require redesign of control measures – Change to operating principles/control mechanisms C. Change in software/interpretation of results – Operating system/database structure/channels of interoperability – User interface – Algorithms/diagnostic features D. Change in materials and manufacturing – Change of equipment components/reagents/calibrators – Ingredient or material from new supplier does not meet existing specification – Change in shelf life

1. The table guides us through initial steps of the risk analysis. It corresponds to the decision step “Change to marketed test” in Figure 2. We list here elements that may or may not carry a hitherto unrecognized risk. A failure mode effect analysis will guide us on the appropriate validation experiment to perform to exclude risk, or to decide on mitigating measures. The residual risk will determine which path we then follow in the decision tree in Figure 2.

3.2 Method validation: fulfillment of the requirements of ANNEX I

LDTs have to comply with the requirements of ANNEX I. The method validation will recognizably address: scientific validity (the measurand is related to the diagnostic purpose), analytical performance (much as in the old directive), clinical performance, and safety. A risk-based analysis has to give evidence of the benefits and safety. The legislator took a “precautionary” approach to safety, naming multiple hazards for patients, users and bystanders. Published summary statements assure transparency for the end-user.

The MDCG developed guidance intended for IVD manufacturers on the implementation of ANNEX I [27], [28], [29]. Guidance is not legally binding, but it provides a backdrop against which to judge due diligence and state-of-the-art implementation. Guidance cannot cover the wide range of clinical applications, analytical methods, and HI-specific implementations of care pathways. With respect to the latter, clinical and operational requirements can set HI-specific performance criteria. The EFLM WG on test evaluation established a comprehensive framework for test evaluation that encompasses five key elements [26]. The EFLM Test Evaluation framework goes beyond the IVDR requirements as besides scientific validity, analytical and clinical performance, also clinical effectiveness, cost-effectiveness and the broader impact of medical tests are considered. A test is fit-for-clinical-purpose when it produces actionable results fulfilling the needs of the specific clinical care pathway. Needs not met by existing devices correspond to the IVDR absence of an equivalent device criterion. Although not explicitly cited as such by the IVDR, actionable results, clinical effectiveness and cost efficiency correspond to clinical benefit in the IVDR benefit-risk ratio evaluation. The EFLM TF on Performance Specifications in Laboratory Medicine also proposed a specific framework for setting analytical and clinical performance specifications for the purpose of method validation and quality assurance [30]. The EFLM WG-ISO/A also proposed guidance for the distinction between method validation and verification according to the ISO 15189 [31]. A corollary of the advent of the IVDR is that from now on a broad segment of compliant commercial tests are validated in a transparent manner, and henceforth method verification by the laboratory suffices for these tests. For modified tests, a risk analysis can identify the elements that need additional validation, without a need for a repeat validation of unaffected aspects. In what follows, we strictly adhere to the IVDR requirements. However, the above-mentioned frameworks provide useful guidance.

3.3 Drawing up summaries for public information and inspection purposes

For all LDTs, the HI makes publicly available the legal identity of the HI, details identifying the device, a self-declaration of conformity with ANNEX I and when applicable a statement on unmet requirements with a reasoned justification. A centralized national depository of these data can assist laboratories in search of a HI for outsourcing tests that they do not perform in-house. The laboratory manual for requesters preferably refers to this depository. Otherwise, the laboratory manual can serve as a tool for dissemination of the requested information.

For class D LDTs, the laboratory draws up documentation that allows the CA to understand the manufacturing facilities and manufacturing process, the intended purpose of the device, fulfillment of safety and performance requirements. In an ISO 15189 compliant quality system this documentation exists and is readily accessible.

3.4 Surveillance plan – maintenance

Bookkeeping of manufacturing and use of devices is required [25]. Throughout the full life cycle of a device, changes in conditions surrounding the use of the devices, in the manufacturing processes, and in safety measures shall be evaluated and documented, and performance data shall be regularly reviewed and updated, and unwanted events and corresponding corrective and preventive actions (CAPA) shall be documented. It forms the basis of continuous vigilance.

4.1 Changes in matrix for well-established methods

The laboratory offers comprehensive metabolic [47] and hematologic [48] panels for the detection and diagnosis of ailments, and for follow-up of treatments. These high-volume applications are vested in a long clinical tradition, as exemplified by their occurrence in multiple clinical guidelines and medicine textbooks. Thus, additional clinical data are not required. The manufacturer validated some of these tests for blood serum as the matrix. Serum is a typical sufficient matrix for laboratories receiving samples from remote locations. Other laboratories receive most of their samples from hospital wards and work with heparinized plasma to shorten the turn-around-time, and the time-to-diagnosis and-treatment, operational requirements of the care pathway. The change of matrix is a critical element that however can be validated in isolation. Samples for both matrices are drawn from subjects covering the whole pathophysiologic range of results (corresponding to required clinical data), and results for the plasma matrix are correlated with the serum comparator. The experiment is covered by the HI ethical committee, and the laboratory has proof of informed consent by the subjects. An appropriate statistical analysis is performed (analytical performance), resulting in appropriate review of reference ranges (clinical performance). Appropriateness is evaluated by comparison with published intra-and inter-individual biological variation, and by participation in PT schemes. This is a wide-spread policy across HIs, and reliance on published and shared information is also acceptable.

A properly executed ISO 15189 audit will cover all of these elements. Strictly speaking, the implemented method qualifies as an LDT, but the ISO-15189 conform validation should suffice as an argument validating that (Figure 2, Table 3) the change was not critical and that additional administrative burdens are not warranted.

4.2 Use of tests outside their standard context

A surgeon seeks advice about the nature of a liquid oozing from a drain. The laboratory reacts to this clinical question by measuring proteins, lipids, metabolites, and organ-specific enzymes, released nucleic acids and a cytologic examination of cells. The clinical question does not turn this into a class D issue. With respect to the quantitative tests, a literature search of case reports shows that these non-validated tests can and have been used to answer the surgeon’s question [49]. This confirms the scientific validity and clinical performance. A full-flagged analytical validation is not commensurate with the intended use, and hence not appropriate. In concordance with published literature the responses that will lead to a conclusion are clearly off-scale. Thus, analytical performance does not require that the exact same tests as in the publications are used and does not require additional proof of traceability, linearity, or measurement uncertainty. These arguments constitute our validation of the tests, and “the non-validated status in ISO 15189 parlance” will be properly communicated with the users.

Changes to the stated intent or the matrix for which the tests are certified turns them into an LDT. The laboratory offers this “off-label” use for occasional compassionate use. In the laboratory manual and on the reports, we define the test stating the matrix and it’s off-label unvalidated nature. We comment that the interpretation relies on a conjecture of test results and needs to be confirmed by the clinical presentation and where possible by appropriate specialized imaging or other examinations. As a class C test, we have not to report this to the CA. In case of inspection by the CA we refer to the laboratory manual and the Laboratory Information System (LIS) report definitions of the tests, which informs the user about the drawbacks cited in the above summary validation.

With respect to the cytologic evaluation, this is essentially an inspection and not a measurement activity [50], and thus, falls beyond IVDR device classification rules and performance criteria. With respect to detection and characterization of released nucleic acids within alternative body fluids (pleural, ascitic, plasma, etc.), when the tissue cannot be analyzed, this requires caution, in the absence of internationally recognized standards in terms of volume of samples, type of molecular assays, cut-offs for reporting genetic variants, actionability of identified targets. It requires skilled laboratory specialists who communicate their expert evidence with the clinician in a deliberation dialogue [22], and this professional service is also beyond the scope of the IVDR.

4.3 Precision diagnostics

The standard of care for solid [51, 52], myelo-[53, 54] and lymphoproliferative [55, 56] neoplasms involves selection of treatment options guided by clinically relevant genomic alterations (CRGA), and this has resulted in spectacular improvements in clinical outcomes. Genetic testing focuses on expression of germline and somatic genetic variants. Two very different diagnostic strategies prevail. Targeted detection of genetic “hotspot” mutations known to be pathogenic using allele-specific methods or next generation sequencing (NGS). Secondly, comprehensive genomic profiling (CGP) can be used to scan for genomic alterations (GA) across many cancer-related genes. Both strategies require their own metrics and pipelines, yet both need to comply with the IVDR.

With the rapid development and availability of massively parallel analytics technologies, all tumors will be increasingly screened for a broad variety of actionable genes without prior knowledge of the clinical utility of the diagnostic results. Specifically, the number of druggable targets and candidate therapeutic compounds rapidly expands for personalized treatment regimens. Hence, we need to acknowledge that the number of targets to be tested and potentially acted upon by the oncologists, is highly dynamic and thus requires an expandable scope definition. The respective diagnostic tests are under continuous development and evaluation in research-oriented laboratories and (international) networks actively advancing translational medical approaches outside the realms of classical “intended-use” applications. In addition, highly selective diagnostic question will determine the use of analytical platforms, growing databases and IT tools within dynamic workflows. So, continuous changes in diagnostic intent and in method design and use defines these in-house diagnostic tests and consequently implies the absence of timely available commercialized alternatives.

Performance qualification and post-market surveillance follow from the “dynamic” appreciation of “clinical performance” in the networks. Continued maintenance of competence referring to this increasing body of knowledge is in line with ISO 15189 accreditation criteria. Sample processing, analysis, and predictive analysis delineated from molecular information involving bioinformatics pipelines and artificial intelligence engines are often spread over multiple participating HIs. Contractual arrangements stipulate the exchange of services, and external quality assessment and assurance. The interpretation of the individual case results from an integrative multi-disciplinary deliberation dialogue involving laboratorians, pathologists and oncologists [57].

Similarly, laboratories use custom gene panels and bioinformatics pipelines for genetic testing of rare diseases, often adapted to the specific medical condition in question in accordance with national or international recommendations. Moreover, there is no systematic harmonization of pipelines for exome sequencing (regularly improved supplier kits) and genome sequencing, and differences between tools are more related to the number of genes covered by the assay than the complete utility in clinical settings.

Mutatis mutandis, as modern medicine moves towards more individualized health care, much of the above also will apply to other fields in the forefront of modern diagnostics, like metabolomics and proteomics [58], that also involves customized platforms, and shared dynamically growing databases and analytical pipelines.

In some countries, laboratories already are required to seek accreditation of their competence for these activities under ISO 15189 [14]. “Flexible” reference to a “broad application scope” and to “ad-hoc builds of analytical infrastructure” is managed compliant with ISO 1589 [15]. A “fixed” static preconceived planning-implementation-cycle is replaced by “dynamic” well-documented “project management” with “continuous surveillance” and registration and correction (CAPA) of undesirable outcomes. Only by applying the same approach “creatively” to IVDR requirements, laboratories will be able to keep serving patient’s needs, simultaneously safeguarding against substandard practices and to retain spearhead translational research within the EEA.

4.4 Companion diagnostics

Precision diagnostics applies to selection of treatment options to optimize expected outcomes and/or minimize undesirable side effects. They can be used both at the time of initial diagnosis, or for follow up in the course of the treatment. Often tissue is collected for future analysis in the course of the disease, and/or deposited in biobanks. When these applications are a prerequisite to qualify for a particular clinical treatment, they are defined as companion diagnostics (CDx). For in-house CDx, the validation of performance is covered by registration of the clinical study protocol and audits by clinical study sponsors [59, 60]. HI often finetune CDx to fit into the HI workflows [61]. The European Medicines Agency (EMA) discusses CDx aspects on a regular basis, foremost analytical and diagnostic prowess [62]. MDCG guidance acknowledges the use of in-house LDTs in medicinal clinical studies [24]. The existing safeguards should suffice to ascertain safety and thoroughness of clinical studies, without impeding access of study sponsors to cooperating HIs.

4.5 Software

The IVDR has broadened the definition of an IVD to include also software, either as an integral part of a device, or as a stand-alone application. In summary [63], the IVDR applies when a programmed algorithm provides information relating to the intended medical purposes of the IVDR (art 2.2) and the input is substantially dependent on IVD’s. The component with highest IVD risk class dictates the ultimate risk class. When the software is developed or sold as a package limited to storage, archival, communication and simple search functionality it does not classify as medical device [28]. This would apply to the LIS, and the middleware used to communicate with the analyzers.

Standard parametrizations of software, like reference limits, conversion of units, but also simple calculated parameters, such as creatinine clearance or metabolite over creatinine ratios are part of the diagnostic interphase that laboratory professionals establish for supporting diagnostic stewardship. Not implementing them should be interpreted as not safeguarding the intended use of the underlying devices. The above simple algorithms should not classify as LDTs in their own right, not warranting additional administrative burdens. Algorithms creating new diagnostic information should be carefully reviewed. Substantial dependence on multiple IVDs or other data sources, or by way of example more complex algorithms used for prediction of risk or of resistance or responsiveness to drugs, or bioinformatics pipelines (Section 4.3) involved in assay set-up and analysis may classify as LDTs. The latter are in some countries already certified under ISO 15189 [14]. Inappropriate variant calling and the misinterpretation of the variant effect enriched by the NGS pipelines can dramatically influence the patients’ treatment. For NGS assays working on tumor tissues, clinical cut-offs of acceptability for tumor variant allele frequency (VAF) (%) or HRD scores are not unequivocally defined, impeding to achieve the goal of responding to overall clinical needs [64]. If the intended use of the comprehensive or targeted NGS panels is related to the treatment, bioinformatics pipelines should use appropriate filtering criteria to discriminate passenger from driver mutations, providing clinically relevant information on the actionability and possible drug treatment of the called-reported variants.

A typical vertical ISO 15189 audit handles this from test request to report, consistent with how the standard and how laboratories treat this as part and parcel of the proper clinical use of the underlying device. Resulting from an overall risk analysis [65], the laboratory has a written policy identifying the need for data integrity, protection against unwanted and malicious loss of data integrity [66], and compliance with the EU General Data Protection Regulation (GDPR) [67], besides regulations on the safe and supervised deployment of big data engines and artificial intelligence [68, 69]. These HI-wide concerns are covered by appropriate service level agreements with the relevant providers. The risk analysis, applying to the complete life-cycle of software [70], summarizes critical changes to the data management infrastructure which need ad-hoc revalidation of the integrity and correct functioning of the system (Table 3).

5.1 About shared liability, due diligence and intended transparency

As discussed in the introduction, the primary intent of the new legislation is to guarantee safe and clinically effective medical tests on the EU-market. Legal requirements and standards cannot cover in absolute detail the wide range of clinical applications, analytical methods, and HI-specific implementations of care pathways. Absolute legal certainty would reduce flexibility, increase costs and stifle innovation, and thus might be detrimental to the patient.

The new legislation allows for creative interpretations as in many instances it calls for judgements on the appropriateness of implementing clauses. All parties involved have to share responsibility. Parties can demonstrate due diligence by appropriately documenting their decisions. The authors call for the profession to claim its role and take up its responsibility. It also calls for the industry to define and demonstrate the intended use of commercial devices at the fullest, and to give due attention to design and software aspects. While a “restricted intended use” certainly reduces validation costs for the industry, it will shift a self-repeating burden to the individual laboratories and also risks to overload the CAs. The manufacturer’s post market surveillance and clinical follow up is an opportunity to identify shortcomings in the intended purpose of devices, and to update their certificates accordingly. By taking up responsibility, manufacturers have an opportunity to stand out between competitors. Finally, we call for the European and national legislators to ascertain lean implementations reducing administrative burden and timely adjustments as dictated by experiences.

There is little awareness of the IVDR across Europe [11]. The authors call upon the profession to educate the lawmakers on the undesirable side effects of a blind stringent implementation of the IVDR. There is a risk for patient and consumer access to in vitro diagnostics (IVDs). Laboratories may become unable to implement all needed LDTs. Also, Europe risks to lose research and development as these activities are likely to move to environments with more lenient regulations.

5.2 ISO 15189 is a sufficient tool for IVDR compliant implementation of LDTs

It is our opinion that the mention of ISO 15189 (IVDR 5 (5) c) elevates it to a harmonized standard. In Table 1, we illustrate the correspondence between elements of the IVDR 5 (5) and clauses in the ISO 15189. In Section 3, we illustrate how standard ISO 15189 conform implementation of LDTs covers these requirements. In Section 4, we illustrate this by way of a selection of special cases. ISO requires concordance with applicable laws. Apparent gaps [17] between IVDR and ISO 15189 can be closed by proper interpretations and corresponding implementation.

Up to now, LDTs were often not presented by the laboratory for incorporation into its ISO 15189 scope, for reasons of absence of standard proficiency testing (PT) schemes and in the case of low volume testing for economic reasons. The ISO 15189 provides for alternatives to PT. With the advent of the IVDR, all LDTs have to conform to the IVDR, and information about such conformity has to be readily available. The laboratory can fulfill this requirement by presenting the LDTs to its AB as a fixed or flexible scope [15]. We favor a flexible scope allowing for generic validation of first-in-class devices. An accompanying technical addendum lists then the LDTs for potential evaluation by the CA.

The CA is competent with respect to the manufacturing process of the devices, the fulfillment of applicable elements of IVDR ANN I, and documentary obligations. Paraphrased, the documentation has to show that a device reliably does what it claims to do. With respect to modifications of existing commercial devices, its competence relates to the argumentation on the need for modifications of the intended use or of the design of the device, the significance of the change and the improved benefit/risk ratio resulting from the modification. Per May 2028, the laboratories have to demonstrate due diligence with respect to the market exploration for viable commercial alternatives.

5.3 Reduction of administrative burden – use of ISO 15189 accreditation mechanisms

For the implementation of the IVDR, national legislators have to install the CAs mostly by assigning new competences to existing bodies, taking into account their own legislative ecosystem and the organization of the healthcare system. Typically, licensing of HI and laboratories, auditing according to ISO standards, and oversight of pharmaceuticals and medical devices are the competence of different administrations. In some countries ISO 15189 accreditation is a prerequisite for licensing of the laboratory, and in others accreditation is voluntary [14]. In some countries professional organizations are self-regulatory [19], while in others all regulatory initiative comes from relevant authorities. The situation may be further compounded by various degrees of defederalization of the health care system [19], with a corresponding hierarchy of competences, and of primacy of regulatory frameworks. By way of example, in Germany, healthcare is regulated by the “föderierte Länder” and cannot be regulated by national bodies. A quality regime (“RiliBÄk”) which essentially mirrors ISO 15189 is mandatory while accreditation to ISO 15189 is voluntary and legally subordinate to RiliBÄk requirements [19]. Obviously, we cannot give a one-size-fits-all advice.

Notwithstanding the above, it is the opinion of the authors that the legislator in interpreting the IVDR has to minimize administrative burdens. Striving for perfection in the legislation might be the enemy of an effective legislation. It is much easier to add elements to a law later, then to scrap superfluous rules. The scope and 2027 end date of IVDR article 111 may come too late to amend unforeseen and unwanted side effects of the regulation. Coordination with other regulations on chemicals and information technology has to be updated in due time.

Under the IVDR (5 (5) g), the national CA can extend its authority to request documentation for non-class D devices. If the national legislator were to make this mandatory, the administrative workload will increase immensely. Also, the CA might be enticed to mimic a NB approach, and ask for notification and approval of planned performance studies prior to their execution. It will add to the administrative burden and delay the putting into use of the LDTs. Also, the CA might ask to report IVDR ANN I compliant safety and performance studies in a pre-determined format. It risks to reduce the validation file to a checklist while not necessarily improving the appropriateness of the validation studies. It is the opinion of the authors that proper risk analysis is a better way to guarantee the appropriateness of the validation files. It requires all parties to share responsibilities (Section 5.1). For the sake of communication with the CA we recommend that the laboratory in its general operating procedures for each element of the validation file refers to the applicable clauses of IVDR ANN I.

Furthermore, and notwithstanding that the CA can at will call for information or do unannounced audits, it would be wise to make maximum use of accreditation programs already in place. The introduction of new regulatory and administrative mechanisms will come with investments of tax payer money. Imposing administrative fees on the HIs corresponds to futile recycling of tax-payer money. The resulting overall opportunity cost for the health care system can be minimized by avoiding duplication of auditing and inspection activities. Solutions will differ by country, as the obligation for medical laboratories to adhere to the ISO 15189 is not universal [14]. In some countries relevant regulatory or oversight authorities are already embedded in AB audits to participate as an observer auditing the applicable specific legal requirements, and/or AB make use of checklists to evaluate specific legally-binding requirements. Embedding the CA into the audit process can be a practical solution. Although laboratories have no choice in it, acceptance of the IVDR will increase when the laboratories can continue to operate in an already existing, familiar and well accepted regulatory space.

5.4 A call for balanced regulation

Calls for additional guidance stem from rightfully perceived incompleteness of the law [71]. Other initiatives are intent on providing generalized approaches to the implementation of the IVDR [16] to avoid every laboratory reinventing the same wheel. As illustrated in Sections 3 and 4, generally applicable guidance will be hard to distill, and the product might draw as much criticism, as the open questions it intends to settle. To be generally applicable MDCG guidance documents should stick to what has to be done, and not to how it has to be done. This also applies to new standards. Thus, these initiatives risk to mostly duplicate existing regulations and guides, without resolving specific issues. Acceptance of additional standards as harmonized standards, may force laboratories to apply for superfluous accreditation. On the other hand, if such initiatives are very specific on the how, then the price we pay for a reduction in liability is the loss of freedom for reasoned optimized implementations and for innovation.

5.5 Evaluation of equivalence of commercial devices

The evaluation of the presence of an equivalent device on the market (Section 3.1.2, Figure 2) requires a reasoned decision on whether marketed devices meet the appropriate level of performance required in the HI’s clinical care pathway (IVDR 5 (5) d). It is the opinion of the authors that this performance criterium should not be restricted to the safety and performance criteria set forth in IVDR ANN I. Non-equivalence refers not only to intended use and predefined performance specifications, but should also be extended to operational, organizational and economic aspects. By way of example of operational considerations, and not limitative, we list here what are also valid elements in the evaluation of public tenders [72]. Implementation of devices according to manufacturer’s instructions may not meet turn-around-times requested in the care-programs serviced by the laboratory, or the laboratory does not have access to particular analytical platforms for some tests. Physical footprints and limited financial space can prohibit the implementation of multiple platforms. Manufacturers may not be able to guarantee continuous supply. Not allowing for these valid extended arguments can force the laboratory to reduce their portfolio and will impede optimal patient care. When opting for an LDT, the laboratory shall transparently document its reasoned justification and resort to evaluation and management of the associated risks.

5.6 Evaluation of critical changes to commercial devices

In Figure 2, we introduced the need for a reasoned decision on the criticality of modifications to commercial devices. The MDCG produced a document with guidance on what it considers significant changes in legacy IVD Medical Devices [27]. While this applies to commercial devices and derogations for the transition period between the directive and the regulation, we believe that it can serve as a guideline for the evaluation of the equivalence requirement of the IVDR (IVDR Art 5 (5) d). This double use would ascertain a level playing field for industry and HIs. In table 3, we recapitulate major elements that require a reasoned decision on their criticality, and where appropriate will have to be addressed while validating the LDT for use. It is the opinion of the authors that the profession can take up its responsibility to decide when a modification is minor and the device can be considered to remain a device as intended by the manufacturer, and when a modification will require validation of safety and/or performance characteristics to the extent appropriate for the planned modification, while relying on the manufacturer’s validation for other aspects.

5.7 Evaluation of home-made LDTs

Commensurate validation applies to modifications to commercial devices (Section 5.6) and to de novo devices developed in the laboratory. The laboratory will resort to its own LDTs mostly, because they fill in niche applications, and there is insufficient interest from suppliers to develop them and to put them into use. By way of an example, we cite here the phenotypic and genotypic identification of long known but rare inherited errors in metabolism. In other cases, niche applications correspond to research interests of the laboratory evolving in innovative translational clinical research. By way of an example, we cite the application of mass spectrometry to the characterization of protein patterns [58]. Obviously, in these examples the original fundamental research will fulfill requirements about scientific validity, and the translational research will document the fitness for clinical use [33]. The validation will have to be supplemented with an analytical validation. A yardstick for the appropriateness of this exercise is whether the documented performance criteria (e. g. traceability and accuracy), will allow future translation to the wider community. Proper peer-review should guarantee this.

The authors believe that the above demonstrates the indispensable role of the medical laboratory professionals in the IVDR roll-out [22]. The IVDR main focus is the manufacturing and putting into use of safe and effective devices, that is prior to the production of individual test results by the laboratory. For the laboratory the manufacture of an LDT device and the laboratory’s core business, the use of the device for the production of reliable test results, is one continuous process. The latter is embedded in the clinical function of the laboratory in the hospital or care setting where it engages in consultative services participating in a deliberation dialogue with users about which tests to request, and how to interpret results. Thus, fulfilment of the IVDR requirements comes as a second nature. This is already going on for some decades, gradually being further refined and broadened, guaranteeing fulfillment of all IVDR requirements (among them clinical evidence).

5.8 Putting into use of LDTs

Also, it is the opinion of the authors that the restriction to non-industrial scale of manufacturing of devices, is covered by non-transferal of devices to other HIs (IVDR 5 (5) a). HIs can in-house analyze outsourced samples from other HIs. HIs are allowed to exchange protocols too, which is relevant in case of consortia. A different interpretation would be detrimental to patient care. Indeed, for highly specialized low volume tests reliable analysis and adequate consultative services require minimum critical volumes. Open market principles of the EU also imply that the exchange of services is not hampered by national borders.

5.9 Surveillance plan – maintenance

By means of a surveillance plan, we propose that the ISO 15189 compliant management review provides in a section for the review of changes in conditions surrounding the use of LDTs, in the manufacturing processes and in safety measures, and in performance data and vigilance data (unwanted events and corresponding CAPA). For the sake of inspections by the competent authority it is recommended that this summary is readily available and searchable. Unwanted events have to be traceable to manufacturing and use data [25].

5.10 Off-label and RUO are not part of the IVDR

The use of devices for scientific research or for clinical studies, while not serving a direct diagnostic purpose, are not within the scope of the IVDR.

5.11 Training in regulatory affairs of professionals

The EFLM recognizes a need for dissipation of information on the IVDR, and for specific training on its implementation with respect to LDTs. The EFLM plans to make content available on its e-learning platform.

5.12 Monopolies and patents

New technology almost invariably will be subject to patents, even when they emanate from research partly published in the public domain. Patents create monopolies, and are counter to free-market price setting. The IVDR protects commercial devices against competition from equivalent home-made devices which profit from a relaxed regulatory environment. However, the commission should monitor potential abuse of dominant positions [73].