AI-Driven Quality Assurance Frameworks for Decentralized Clinical Trials

Abstract

Traditional quality assurance (QA) processes used in electronic Clinical Outcome Assessment (eCOA), and Decentralized Clinical Trial (DCT) platforms rely on a variety of manual, reactive approaches with many regulatory risk elements. The purpose of this research is to introduce an AI-powered Predictive Quality Assurance (PQA) system that uses Ensemble Machine Learning Models to predict problems with the DCT platform or eCOA application prior to their release into production. Using simulated and anonymized QA data sets, the PQA system will implement Supervised Learning, Deep Learning and Hybrid models to generate Risk Scores at the Module Level as well as Optimize Regression Test Coverage in Real Time. The evaluation outcomes showed improvements in defect prediction accuracy and reductions in Defect Leakage post Release, and thus the use of AI-Driven QA can enhance the dependability of software applications, prepare the software applications for Regulatory Audits, and help enable the Digital Transformation of Clinical Research Systems.

Ähnliche Arbeiten