Thesis Format
Monograph
Degree
Master of Science
Program
Computer Science
Supervisor
Dr. Kamran Sedig
Abstract
Adverse drug events (ADEs) pose a significant challenge in pharmacovigilance due to their complex, multidimensional nature and the limitations of static visualization methods. This thesis presents ENVISION—OpenFDA Adverse Drug Events Database Exploratory Interactive Visualization—an interactive system that transforms raw ADE data into visual insights. ENVISION addresses the important need of generating user friendly visualization of data by combining the hierarchical visualizations that utilize a tree map and Sankey diagram within it and real-time filtering and drill down capabilities. In this work, we use a representative openFDA ADE dataset as our project and it uses human–computer interaction principles and web technologies (e.g., D3.js and JavaScript) to visualize intricate relationships between patient demographics and drug information with reported outcomes. This system facilitates efficient hypothesis generation and informed decision-making by enabling exploration of ADE data across multiple dimensions. Ultimately, ENVISION enhances data comprehension and supports effective drug safety monitoring in dynamic healthcare environments. This interactive system paves the way for safer healthcare.
Summary for Lay Audience
This thesis explores how interactive visual tools can make complex data about ADEs understand and use the healthcare easier. ADEs are that harmful effects which occurs on patients taking medications which then lead to serious consequences like hospitalizations or even death. Organizations like the FDA collect these data traditionally, but due to the volume and complexity of the reports they are difficult to analyze using simple, static reports.
However, it is challenging to deal with the complexity of the ADE data structure, the research meets this challenge with introducing an interactive system called ENVISION which takes raw ADE data and transforms it into clear, dynamic visual representations. ENVISION uses interactive elements like hierarchical treemap and Sankey diagram. Reports are grouped by country and drug in treemap and are denoted by sizes and colors where number of reports and type of outcomes observed (e.g. death or not serious) are indicated. Users can now easily identify trends, for instance, which countries report of more events or which drugs are related to serious outcomes.
This can be complemented by the Sankey diagram showing the connection of a selected drug with both its medical uses (indications) and the adverse events that have been reported. Connecting lines of varying thickness and color indicate the relative frequency of each link, and show users at a glance how many drug-use combinations may be a problem.
Interactive filters like sliders for age, weight, and time, as well as search functions that make it possible for users to refine the data and narrow their search to the aspects that interest them are all part of the system too. This dynamic exploration drilles both experts and non-specialists down from a broad overview to detailed information required to identify safety signals and understand patterns.
The results indicate that ENVISION is effective not only in simplifying complex data so as to improve understanding of the information but also to accelerate the review of important trends. With this real time accessible data, decision on drug safety monitoring can be made more effectively. Overall, this work demonstrates that well-designed, interactive visualizations can bridge the gap between overwhelming amounts of data and the practical insights needed to protect patient health, offering a replicable model for future advances in pharmacovigilance.
Recommended Citation
Jain, Sanyam, "Exploring Adverse Drug Events Through Interactive Visualizations" (2025). Electronic Thesis and Dissertation Repository. 10696.
https://ir.lib.uwo.ca/etd/10696
