Article type
Year
Abstract
Objective: To help participants navigate through and abstract high quality data from clinical study reports (CSRs), an important source of unpublished data for systematic reviews. We will share our experience and discuss challenges encountered and our working solutions.
Description: The structure and content of CSRs are laid out by the International Conference of Harmonisation of Technical Requirements of Pharmaceuticals for Human Use. A CSR is a detailed description of a single study in which the clinical and statistical methods, results, and conclusions are integrated into one report. Recent data sharing policies and high profile campaigns for the inclusion of all data in systematic reviews have resulted in increased access to CSRs, yielding a plethora of information that is not always available in published journal articles. However, CSR data abstraction (DA) can be challenging due to an overwhelming amount of data organized in unfamiliar ways. For example, a CSR may be as long as 3000 pages with as many as 200 outcomes and 700 adverse events (AEs), while a published report on the same trial may only have 6 outcomes and 15 AEs. After a brief introduction, we will guide participants through the partial abstraction of a CSR by setting up a DA 'scavenger hunt' where teams will work through different steps of the CSR DA process. We will provide guidance to achieve high-quality and efficient CSR DA. We will end with discussion, allowing participants to share their experience.
Description: The structure and content of CSRs are laid out by the International Conference of Harmonisation of Technical Requirements of Pharmaceuticals for Human Use. A CSR is a detailed description of a single study in which the clinical and statistical methods, results, and conclusions are integrated into one report. Recent data sharing policies and high profile campaigns for the inclusion of all data in systematic reviews have resulted in increased access to CSRs, yielding a plethora of information that is not always available in published journal articles. However, CSR data abstraction (DA) can be challenging due to an overwhelming amount of data organized in unfamiliar ways. For example, a CSR may be as long as 3000 pages with as many as 200 outcomes and 700 adverse events (AEs), while a published report on the same trial may only have 6 outcomes and 15 AEs. After a brief introduction, we will guide participants through the partial abstraction of a CSR by setting up a DA 'scavenger hunt' where teams will work through different steps of the CSR DA process. We will provide guidance to achieve high-quality and efficient CSR DA. We will end with discussion, allowing participants to share their experience.