Obtaining maximum value from disparate patient data sources through consolidation and harmonization

Data generated by clinical trials are spread across many repositories with many different formats, making it difficult to access all available insights outside of a particular study or trial. Such data may take the form of imaging data, bio samples, genomic data, demographic information, adverse events, etc. These data often remain within one clinical trial and are never used again.

Traditionally, it requires a lot of manual effort to collect all the data to allow scientists to search for relevant data assets in such a landscape. This collection is usually carried out by a data scientist in response to a researcher’s request and can take weeks or even months. However, there is an immense value if these data can be reused effectively, potentially avoiding repeating expensive studies. If the patient has given the correct consent, their data could be used for research into similar drugs or diseases.

The outcome 

ONTOFORCE’s DISQOVER platform was chosen because it allows data consolidation through a knowledge graph that lets people explore the information space directly, rather than confining them to a traditional, limited search format. Data from disparate sources and formats, both publicly available and generated in-house by our customer, can be seamlessly combined with a user-friendly search function to empower researchers to find suitable data, quickly and easily, for exploratory research. Moreover, the platform is not a fixed data model: it is adaptable and flexible. The model can be extended to seamlessly integrate any new data that may become available, for example proteomics data.

Our customer has implemented the initial use of ONTOFORCE’s DISQOVER to enhance their data integration and is now empowering its researchers to generate new insights. Previously, the only option was to search through clinical repositories to request potentially relevant data. This could take six months to generate enough information to build confidence in a hypothesis. Thanks to DISQOVER, this can be done in hours.

Furthermore, DISQOVER can reveal new data that previously never would have come to light. As our customer’s Data Strategy Specialist explained:

“A researcher developing a model to predict an adverse event during chemotherapy was aware of three or four studies with interesting patients. With improved searches, he could identify another five studies in different therapeutic areas which contained relevant patients. So, his pool of subjects or data for hypothesis building could be threefold larger than initially thought.”