Artificial intelligence, in particular machine learning (ML), has emerged as a key promising pillar to overcome the high failure rate in drug development. Here, we present a primer on the ML algorithms most commonly used in drug discovery and development. We also list possible data sources, describe good practices for ML model development and validation, and share a reproducible example. A companion article will summarize applications of ML in drug discovery, drug development, and postapproval phase.

Introduction Patient-level data sharing has the potential to significantly impact the lives of patients by optimizing and improving the medical product development process. In the product development setting, successful data sharing is defined as data sharing that is actionable and facilitates decision making during the development and review of medical products. This often occurs through the creation of new product development tools or methodologies, such as novel clinical trial design and enrichment strategies, predictive pre-clinical and clinical models, clinical trial simulation tools, biomarkers, and clinical outcomes assessments, and more. Methods To be successful, extensive partnerships must be established between all relevant stakeholders, including industry, academia, research institutes and societies, patient-advocacy groups, and governmental agencies, and a neutral third-party convening organization that can provide a pre-competitive space for data sharing to occur. Conclusions Data sharing focused on identified regulatory deliverables that improve the medical product development process encounters significant challenges that are not seen with data sharing aimed at advancing clinical decision making and requires the commitment of all stakeholders. Regulatory data sharing challenges and solutions, as well as multiple examples of previous successful data sharing initiatives are presented and discussed in the context of medical product development.