Project case: Clinical Nutrition Data Warehouse


One of our clients specialises in research and development of nutritional products. Throughout decades of research, a vast amount of data has been acquired for numerous studies. Within each study, data was collected with the purpose of answering specific research questions. Some study aspects, such as study product, study population, overlap with others.

So, what if we were to collect and combine the data of all these studies? Can we verify previous results with larger populations? Can we discover new trends? And can we discover new research questions for future studies?

In 2016 OCS Life Sciences was commissioned to build a data warehouse from numerous nutrition studies. As all studies had used different case report forms and had differently structured source datasets, the data required standardisation. Hence, our client agreed to our advice to apply CDISC’s SDTM guidelines. The project team at OCS was originally led by Lieke Gijsbers and later by Leah van der Meer, both statistical programmers at OCS Life Sciences and having a background in nutrition science and a solid knowledge of SDTM guidelines. OCS performed the data conversion using SAS software, and developed a set of macros that could be used across all studies to ensure equal data derivations across studies.

The creation of the data warehouse was truly a combined effort of OCS and its client. The project team at OCS identified challenges in mapping source data to SDTM and proposed one or multiple solutions. These were then discussed with the client..One of the challenges originates in the fact that SDTM was not designed for nutrition studies, and the studies were not designed to fit SDTM datasets. On the other hand, the data warehouse was not meant for submission to regulatory authorities. Therefore, OCS offered a tailored approach where source data was fitted for SDTM-like domains, following conventions specially developed for the needs of our client.

In 2020 the OCS project team was pleased to deliver the complete data warehouse, containing data of nearly 7.500 subjects and covering 23 SDTM domains. The client now successfully uses the data warehouse to perform new analyses and to answer new research questions.

Altogether we can state that our collaboration has been fruitful and carries on in new projects.

Are you interested in what we can do for you? Contact us by clicking on the "Get in touch" button below this article.