Work Package 2 evaluates how evidence should be generated for innovative medical devices during first-in-human studies and early clinical trials. New methods of generating evidence will be considered and methods for evaluating high-risk devices for children will be assessed. Guidance for the evaluation of artificial intelligence in medical devices and standalone software as a medical device will be reviewed and developed.

Objectives

Icon entangling difficult situations

To review and recommend alternative designs of clinical studies that can be used to provide high-quality clinical evidence for new high-risk medical devices

Doctor looking at data

To conduct feasibility studies of using the IDEAL framework in the early stages of device clinical testing

Digital brain icon

To survey existing guidance and recommend criteria for the clinical evaluation and regulatory conformity assessment of artificial intelligence and machine learning in high-risk medical devices

Doctor examining child

To review and recommend methods for evaluating high-risk medical devices used in children

Tasks

The Under the leadership of the Department of Cardiology, Bern University Hospital, CORE-MD partners work on a protocol for a systematic review to critically evaluate the methodologies applied in clinical investigations to evaluate high-risk medical devices in cardiovascular medicine.

Read the news

The aims of this task (2.2) are to assess the current standards when conducting randomized clinical trials (RCTs) of medical devices and to report on alternative study designs that, building on the concept of randomization, may allow more efficient use of resources without affecting the quality of evidence that is generated. When comparing different interventions, RCTs are indeed considered the gold standard method for assessing the unbiased treatment efficacy/performance and safety of an intervention. However, RCTs may suffer from limitations such as excessive costs, and the inclusion of highly selected patient cohorts, eventually leading to limited generalizability of their results in real-life
clinical practice. To overcome these limitations, the concept of “pragmatism” in the design and conduct of RCTs has gained more and more attention. Pragmatism in an RCT is meant as a way of disentangling the burdensome structure of conventional RCTs, largely simplifying the inclusion, retention of study participants, and collection of data during follow-up.

Against the backdrop of a multitude of entities (both official and unofficial) formulating suggestions for the use of AI tools in clinical practice, but mostly so at a high level, this task sets out to formulate practical suggestions for the clinical evaluation of such tools. To this
effect 2 phases were initiated. 

The first phase set out to define the principles underlying the clinical evaluation of AI tools. This phase I was executed by organising several (online) meetings with leading experts from the consortium and concluded with an internal document providing the necessary background and detail for these principles.A manuscript reviewing definitions, expert recommendations and regulatory initiatives was prepared and submitted for publication in Expert Review of Medical Devices in October 2022. 

The second phase of the task will consist in the production of a second manuscript providing a more practical guidance for clinical evaluation of AI tools and will contain 4 chapters: risk base approach, matrix of evaluation, ex-ante (pre-market) evaluation and post-market evaluation and surveillance.

 

The aims of Task 2.4 are to review existing published reports on the clinical investigation of high-risk medical devices in children in order to identify and describe methodologies applied in this research area, and to organize a paediatric expert workshop to develop recommendations for clinical evaluation concerning high-risk medical devices in children.

This task is led by the European Academy of Paediatrics (EAP)while the task coordination is done by Stiftung Kindergesundheit (Child Health Foundation) at the LMU in Munich. To establish the Paediatric Expert Panel, the potentially suitable paediatric experts have been identified. A preliminary literature search was performed in the field of paediatric medical devices and their regulatory challenges to get a first overview on the topic.