Before a high-risk medical device is approved for implantation into patients, there should be evidence not only of its performance and safety with a favourable benefit-risk ratio, but also of its clinical efficacy. Regulatory guidance on study methodologies is lacking, however, so the European Commission funded the CORE–MD project (Coordinating Research and Evidence for Medical Devices) to advise regulators on appropriate designs for clinical trials of high-risk devices. The CORE–MD consortium recommends that evaluation should be planned in four stages. Randomised controlled trials should be performed more often, against active comparators reflecting the best available treatment, or using sham interventions with ethical safeguards. Large trials can be managed efficiently using an electronic database or registry. Non-randomised clinical studies can apply objective performance criteria or other validated patient-relevant outcome measures, with adjustments to minimise bias. Full transparency of results from clinical investigations is essential. Proportionate regulation of breakthrough or orphan devices for independently-defined serious unmet needs may involve approval with less evidence, but on condition of subsequent confirmatory studies. These CORE–MD consensus proposals have been submitted to European Union medical device regulators, to be considered as a basis for more transparent and predictable requirements for clinical evidence. Funding: The CORE–MD project was funded as a Coordination and Support action from the European Union Horizon 2020 research and innovation programme, under grant agreement 965246.

Background and purpose — Pooling data on the performance of total knee (TK) implants across registries is only possible if the same TK implant is used across multiple registries and if used in patients with similar characteristics. We assessed to what extent specific TK implants: (i) are used across multiple registries or only in a single registry; and (ii) differ in patient characteristics between registries. Methods — All primary TK implants implanted between January 2020 and December 2021 in the Danish, Dutch, German, and Italian registries were included. We determined the number of registries using a specific TK implant (based on combined femoral-tibial component brand name and fixation/congruency/mobile bearing insert/patella usage). Patient characteristics (age/body mass index [BMI]/sex/ diagnosis osteoarthritis) were compared across registries for TK implants used in ≥ 2 registries ≥ 100 times. Results — 813 different TK implants (577,351 procedures) were used across the 4 registries, of which 53 TK implants (7%) were used in 1 registry (8,000 procedures). 760 different TK implants (569,351 procedures; 99%) were used in ≥ 2 registries of which 47 different TK implants (393,954 procedures; 68%) were used in ≥ 2 registries and ≥ 100 times. Statistically and clinically significant differences in age for the same TK implant across registries were observed for 29 TK implants (62%) and 3 TK implants (6%), respectively; for other characteristics these were for BMI 30 (64%) and 0 (0%) TK implants; for male proportion 23 (49%) and 17 (36%) TK implants; and for diagnosis of osteoarthritis 42 (89%) and 34 (72%) TK implants, respectively. Conclusion — Most specific TK implants and TK procedures were used across multiple registries, but they were often used in patients with different characteristics. This has an impact on comparing implant performances between registries.

Background: The European Union Medical Device Regulation (MDR) requires manufacturers to undertake post-market clinical follow-up (PMCF) to assess the safety and performance of their devices following approval and Conformité Européenne (CE) marking. The quality and reliability of device registries for this Regulation have not been reported. As part of the Coordinating Research and Evidence for Medical Devices (CORE-MD) project, we identified and reviewed European cardiovascular and orthopaedic registries to assess their structures, methods, and suitability as data sources for regulatory purposes. Methods: Regional, national and multi-country European cardiovascular (coronary stents and valve repair/replacement) and orthopaedic (hip/knee prostheses) registries were identified using a systematic literature search. Annual reports, peer-reviewed publications, and websites were reviewed to extract publicly available information for 33 items related to structure and methodology in six domains and also for reported outcomes. Results: Of the 20 cardiovascular and 26 orthopaedic registries fulfilling eligibility criteria, a median of 33% (IQR: 14%-71%) items for cardiovascular and 60% (IQR: 28%-100%) items for orthopaedic registries were reported, with large variation across domains. For instance, no cardiovascular and 16 (62%) orthopaedic registries reported patient/ procedure-level completeness. No cardiovascular and 5 (19%) orthopaedic registries reported outlier performances of devices, but each with a different outlier definition. There was large heterogeneity in reporting on items, outcomes, definitions of outcomes, and follow-up durations. Conclusion: European cardiovascular and orthopaedic device registries could improve their potential as data sources for regulatory purposes by reaching consensus on standardised reporting of structural and methodological characteristics to judge the quality of the evidence as well as outcomes.