Today you can download Johannes Ulander’s presentation and Kirsten Langendorf’s presentation from the PhUSE conference.
SDTM Domains by Query – is it possible?
Johannes Ulander had a packed room for his presentation on SDTM and linked data graphs. In fact, graphs and linked data were a ‘hot topic’ at this year’s PhUSE Conference. As the life sciences industry looks towards improving processes and, perhaps more importantly, improving the quality of the data, graphs seem to provide the solution. Johannes discussed the way that the industry currently spends painstaking hours mapping data from one system to another and the inherent difficulties of eliminating errors in such a system.
Within the pharmaceutical industry, there is a growing interest in trying to find new ways of working with our clinical data, as it seems that we always end up in the same situations. We have multiple mappings of one data point from one data structure to another, with a constant need to express the relationship between data points and its metadata. We are constantly searching for the perfect relational database model, so that the small subtle changes in our collected data can be handled without interfering with the overall model. Maintaining our standards and releasing new versions is very complex, and when we are ready to release a new version of a standard, it is probably already outdated compared with what needs to be collected.
Other industries, e.g. companies creating social network apps for your mobile phone, usually release new apps every other week. And more interestingly, you usually never notice any difference although something must have happened. These companies use linked data and graph databases to a very high degree, which has triggered Johannes’ curiosity and need to understand if there is something that we in the pharmaceutical industry can learn from this technology.
He gave examples using the CDISC SDTM and SDTMIG to represent tables in relational databases, as it is the most commonly used and understood tabular structure for using clinical data, but the discussions is at a general level on the differences between relational databases and linked data.
Johannes is an expert in CDISC Standards, particularly SDTM and ADaM and graph databases. If you want to know more about the benefits of linked data then please take a look at the Glandon Suite MDR developed by our sister company, A3 Informatics.
You can download Johannes’ slides and full paper here:
Define.xml tools supporting SEND/SDTM data process – Using SAS® Clinical Standards Toolkit
Kirsten’s presentation demonstrated a set of define.xml tools developed for a customer (Pharma Company). She showed how the tools create and compare define.xml and are used to specify what data to receive, description of the actual data received and a way to compare the two. The tools were developed using SAS® 9.4 and SAS Clinical Standards Toolkit (CST) 1.7. Her presentation provided an overview of how the define.xml tools are used in a SEND data process. The three define.xml tools were described and demonstrated. To support the maintenance of the CDISC data standards and the CDISC controlled terminology, two tools have been developed to update the SAS Clinical Toolkit’s Global Library of standards.
Define.xml is often a file that is made before data is submitted to the Regulatory Agencies and intended for the data reviewer. However, Define.xml can be used throughout the data collection and reporting process as a tool to:
- Provide specifications to a data provider, e.g. lab or CRO
- Provide ability to compare received data against specified
- Provide data description for internal users of the data, e.g. Data Reviewers and Statisticians
- Data archiving, serving as a ‘table of contents’ of a company’s data
The tools described and demonstrated in the presentation have been developed for a customer to support a SEND data process, but the tools are also applicable for SDTM.
Kirsten is S-cubed’s Principal Consultant with a particular focus on CDISC Standards. She is also a subject matter expert for A3 Informatics. If you want to discuss implementation of the CDISC standards you can contact us here.