Harnessing the power of multimodal big data in pharmaceutical R&D

Abstract

In the face of a mounting R&D productivity crisis affecting the pharmaceutical industry, data and digital technologies are increasingly seen as a potentially transformative avenue in shortening the time required to develop treatments. Large pharmaceutical companies complete between 50 and 300 randomized clinical trials (RCT) every year, with the totality of historical RCT data assets representing an underexplored trove of potential novel insights. Secondary use of RCT data at scale is, however, coupled with numerous challenges, with organizational data siloes, data harmonization and governance issues, privacy concerns, and lack of technological maturity being just some of the challenges. That often makes leveraging the full potential of this rich data source prohibitively complex. To address this challenge, Novartis created the data42 program, an initiative that brought together data from 2,500 historical trials run by the organization, resulting in a pool of clinical and associated biomarker data that cumulatively amounts to 3 million patient years of observations. This talk will briefly outline solutions – technical and governance-related – that were developed to support building and democratizing access to the resulting pool. To illustrate the potential of such a resource, an overview of the impact created by the program will subsequently be provided, including prospective thoughts about potential use cases, such as virtual proofs of concept, external control arms, and large-scale biomarker analyses. It is hoped this overview will serve to demonstrate how digital technologies can make a fundamental impact in shortening the path to new treatments.