The Credibility of Medical Device In Silico Clinical Trials

Computational models of patients and medical devices can be combined in virtual clinical trials — known as in silico clinical trials (ISCTs) — to assess questions related to device safety and effectiveness across the total product lifecycle. By informing device preclinical and clinical assessments, ISCTs can accelerate product development and potentially reduce human subjects in clinical trials. However, before an ISCT is used to inform decisions that can affect human health, its credibility must be evaluated.

In "Toward Trustworthy Medical Device In Silico Clinical Trials: A Hierarchical Framework for Establishing Credibility and Strategies for Overcoming Key Challenges," Exponent's Steve Kreuzer and co-authors, including Food and Drug Administration regulators and leading clinicians and academics, address challenges and strategies for establishing ISCT credibility.

ISCTs represent a generational advancement in the application of computing technologies for medical devices, allowing organizations to scale up methods traditionally used in early device development to later stage device evaluation, enabling more accurate prediction of device reliability and performance. The strategies required to unlock the promise of these methods create unique challenges, most importantly the integration of patient, device, device-patient interaction, and clinical procedure submodels. These submodels may constitute different model types, and therefore combining these into virtual cohorts of multiple patients and mapping acute procedural results to long-term clinical quantities of interest or endpoints necessitates a range of strategies and approaches for generating credibility evidence. 

The authors describe a hierarchical approach to establish credibility of the overall model (i.e., the full ISCT) that involves collecting credibility evidence for each submodel and organizing the evidence into one of eight categories described in recent FDA Guidance. Ultimately that credibility evidence will stem from a variety of sources, including benchtop testing and analysis of clinical data, placing an emphasis on strategic planning to achieve business and regulatory objectives. 

From the publication: "An ISCT is a virtual representation of the real world that needs to be shown to be credible before being relied upon to make decisions that have the potential to cause patient harm. There are many challenges to establishing ISCT credibility. In this study, we begin to address some of these challenges and to identify general strategies for overcoming them."