Clinical Trial Drug Safety Assessment With Interactive Visual Analytics

In this article, we provide guidance on how statisticians can use interactive visual analytics to assist medical personnel through the entire clinical safety review process. For the general assessment of safety data (e.g., adverse events, laboratory measurements), we recommend a review flow with the first step using a display that leverages statistical methods to identify events with stronger evidence of a treatment difference. In combination with clinical knowledge, reviewers can identify events that need further scrutiny. Next, clinical reviewers will be provided with displays that show additional details on these events or patient level information to aid their decision making. For safety topics of interest (e.g., suicidal ideation and behavior), we propose a tailored approach to each topic both at the summary level as well as the patient level. Displays will show only information relevant to the specific topic and for the clinical questions of interest. We also discuss some challenges to fully implement and leverage interactive displays and tools. We encourage broad use of interactive visual analytics in the analysis and display of clinical data. Study sponsor: Eli Lilly and Company.