Date of Award

May 2014

Degree Type


Degree Name

Doctor of Philosophy


Biomedical and Health Informatics

First Advisor

Susan McRoy

Second Advisor

Hong Yu

Committee Members

Rashmi Prasad, Peter Tonellato, Steven Belknap


Adverse Drug Event Detection, Natural Language Processing, Patient Communication, Summarization, Text Mining


Adverse drug events (ADEs) are injuries resulting from a medical intervention related to a drug. ADEs are responsible for nearly 20% of all the adverse events that occur in hospitalized patients. ADEs have been shown to increase the cost of health care and the length of stays in hospital. Therefore, detecting and preventing ADEs for pharmacovigilance is an important task that can improve the quality of health care and reduce the cost in a hospital setting. In this dissertation, we focus on the development of ADEtector, a system that identifies ADEs and medication information from electronic medical records and the FDA Adverse Event Reporting System reports. The ADEtector system employs novel natural language processing approaches for ADE detection and provides a user interface to display ADE information. The ADEtector employs machine learning techniques to automatically processes the narrative text and identify the adverse event (AE) and medication entities that appear in that narrative text. The system will analyze the entities recognized to infer the causal relation that exists between AEs and medications by automating the elements of Naranjo score using knowledge and rule based approaches. The Naranjo Adverse Drug Reaction Probability Scale is a validated tool for finding the causality of a drug induced adverse event or ADE. The scale calculates the likelihood of an adverse event related to drugs based on a list of weighted questions. The ADEtector also presents the user with evidence for ADEs by extracting figures that contain ADE related information from biomedical literature. A brief summary is generated for each of the figures that are extracted to help users better comprehend the figure. This will further enhance the user experience in understanding the ADE information better. The ADEtector also helps patients better understand the narrative text by recognizing complex medical jargon and abbreviations that appear in the text and providing definitions and explanations for them from external knowledge resources. This system could help clinicians and researchers in discovering novel ADEs and drug relations and also hypothesize new research questions within the ADE domain.