WHITE PAPER
Applied Adaptive Design Using Subgroup Identification and Machine Learning
An innovative approach from IQVIA to maximize the chance of phase 3 study success
Jul 13, 2023

The latest artificial intelligence/machine learning (AI/ML) tools leverage phase 2 study data to optimize phase 3 design. Current tools offer an array of design options, statistical models and simulations, enabling rapid exploration and development of novel trial designs. Approaches include modeling and simulating phase 3 patient accrual to determine appropriate patient sample size and study timeline. This document intends to outline how the motivations of Phase-3 design for widest possible label, shortest trial time and smallest possible sample size can be accomplished using the latest AI/ML and adaptive trial design and accrual tools.

Adaptive designs – defined as clinical trial designs that allow for prospectively planned modifications to the design based on accumulating data from subjects in the trial – have potential to reduce a trial's resource requirements and completion time, while increasing the probability of study success. These designs have particular advantages in phase 1 and phase 2 studies, which evaluate efficacy and monitor safety events. Artificial intelligence and machine learning (AI/ML) tools can inform phase 3 design decisions when applied to phase 2 study data, including adaptation of patient inclusion/exclusion criteria for the broadest possible label and optimizing sample size and timeline without compromising study outcomes. Adaptive accrual is also a promising approach, incorporating actual treatment response and hazard rates for efficient trial execution without sacrificing power. Authored by experts from IQVIA, this paper examines the use of adaptive approaches to phase 3 trial design, with potential to reduce cost and time to run trial, while increasing the probability of study success.

Related solutions

Contact Us