Towards Large-Scale Drug Safety Surveillance: A Big Data Perspective

October 4-7, Chicago, Illinois, USA
Tutorial in ICHI 2016

Abstract:
Adverse drug reaction (ADR) is one of the major causes of failure in drug development. Accurate prediction and identification of potential ADRs are required in the entire life cycle of a drug, including early stages of drug design, different phases of clinical trials, and post-marketing surveillance. However, each approach has its unique limitations. For example, non-clinical experiments conducted to study drug toxicology is known to have the translational barrier in terms of its effect in human beings; pre-approval clinical trials are limited by small sample size and cannot predict many adverse drug reactions that are observed in real-world patient populations; whereas post-marketing surveillance methods may be biased and inherently leave patients at risk until sufficient clinical evidence has been granted. These well-recognized limitations inherent to the type and diversity of data sources employed in routine drug safety surveillance, along with increased public concern over the safe use of drugs, have stimulated several worldwide research and legislative initiatives with the objective of improving drug safety surveillance.

On the one hand, a number of emerging computational methods, ranging from data mining, statistical modeling to text mining, has become one of the main driving factors to advance drug safety surveillance. On the other hand, it is widely accepted that next progress in drug safety surveillance depends on a comprehensive approach that examines ADR-related information from a diverse set of potentially complementing data sources, such as combining chemical, biological, pharmacological, large-scale observational health and social network data to predict ADRs in both individual patients and global populations.  In this tutorial, we provide a review of the publicly available large-scale databases relevant to drug safety surveillance, describe the kinds of computational methods that can be applied to them, discuss challenges such as how to evaluate the effectiveness of methods and how to appraise the precise value of each data source, and finally represent opportunities for future work. 

The tutorial targets all healthcare informatics researchers and practitioners who are interested in developing and applying advanced computational methods for drug safety surveillance. No domain knowledge of drug safety surveillance is required. General knowledge of statistics is assumed.

The tutorial aims to benefit participants for:
Understanding recent advances in computational methods for drug safety surveillance
Exploring benchmark data from real-world applications for better evaluation of computational methods
Addressing the current challenges in large-scale drug safety surveillance

Tutors' short bio:
Ying Li is a Postdoctoral Researcher at Center for Computational Health, IBM T.J. Watson Research Center. Dr. Li graduated with PhD in Biomedical Informatics at Columbia University on June, 2015. Her advisor was Prof. Carol Friedman. She also worked as an intern at GlaxoSmithKline during the summer of 2012. Her dissertation research focuses on post-marketing drug surveillance utilizing data mining methods based on observational healthcare data and spontaneous reporting systems (SRS) data. Her other research interests include drug repurposing, text mining and data mining based on Electronic Health Records (EHRs). She has published 8 articles in referred journals and conferences, including Nature Biotechnology, JAMIA, Drug Safety, AMIA and IHI. More details in http://researcher.watson.ibm.com/researcher/view.php?person=us-liying.

Ping Zhang is a Research Staff Member at Center for Computational Health, IBM T. J. Watson Research Center. He is leading translational informatics research in IBM. His research focuses on Machine Learning, Data Mining, and their applications to Drug Discovery and Health Informatics. He has published more than 25 articles in refereed journals and conferences, including AMIA, BIBM, ECML/PKDD, KDD, WWW conferences, and BMC Bioinformatics, JAMIA, Nucleic Acids Research, Proteome Science, Scientific Reports journals. Dr. Zhang served on the program committees of leading international conferences including KDD, IJCAI, UAI, SDM, BIBM, and ICHI. He also serves on editorial board of CPT: Pharmacometrics & Systems Pharmacology, an official journal of the American Society for Clinical Pharmacology. More details in http://researcher.watson.ibm.com/researcher/view.php?person=us-pzhang.