Visualize and Understand Your Safety Data: A Library of Graphical Approaches
FDA, Industry, and Academia Working Group on Visualizing Safety Data
Do you want to convey your clinical development program’s safety information in a visual and intuitive manner? You can at CTSpedia, a website composed of graphs that answer common clinical trial safety questions. You also can learn about best practices for constructing graphs and conveying the data’s key messages.
The graphical approaches featured on the website are a byproduct of a three-year venture by a collaborative working group comprised of experts from the FDA, pharmaceutical industry, and academia. To represent the broad spectrum of safety information collected in a clinical trial, the following three topic areas were considered:
- Electrocardiograms (ECGs) – graphical approaches related to the assessment of ECG data
- Labs/Liver – graphical visualizations to assess laboratory data with a focus on liver function
- General Adverse Event – graphical approaches to view general adverse event information such as data coded using the MedDRA (Medical Dictionary for Regulatory Activities) dictionary
Leaders: Mat Soukup, FDA; Ken Koury, Merck Research Laboratories; Brenda Crowe, Eli Lilly and Company
Members: Rich Anziano, Pfizer Inc.; Mary Banach, Vanderbilt University; Fabrice Bancken, Novartis; Andreas Brueckner, Novartis; Susan Duke, GlaxoSmithKline; Rich Forshee, FDA; Robert (Mac) Gordon, Janssen Research & Development; Larry Gould, Merck Research Laboratories; Frank Harrell, Vanderbilt University; Liping Huang, CSL Behring; Qi Jiang, Amgen; Stephine Keeton, PPD; Andreas Krause, Actelion Pharmaceuticals Ltd.
Each of the groups identified or developed graphical approaches that addressed key clinical questions in the domain of interest. These graphical approaches were reviewed by a fourth group charged with applying good graphical principles and identifying best practices for graph construction. Reproducible working examples of the graphics are included.
Examples include a variety of graphical methods, ranging from complex (multi-panel dot plots) to simple (scatterplots) displays. Users are encouraged to add their perspectives and contribute new graphical visualizations on this wiki.
Resources also are provided for selecting the most appropriate graph type for a given situation. The inclusion of worked examples with sample data and programming code can reduce the resources required to implement graphical approaches.