Clinical Metadata at Takeda

For Julia Fox, Director, Clinical Metadata at Takeda, her role is to define the technical space around clinical metadata and standards and establish processes and tools to facilitate user engagement. Fox’s key responsibilities are stakeholder engagement via outreach and communication, primarily focused on user requirements related to key clinical metadata and standards in end-to-end clinical data processes. The team aims to define and refine technical needs for metadata management, accessibility, and cross-system configurations. We are endeavoring to develop a modern vision for metadata management and usability within its clinical data ecosystem, and thus an aspect of my role is to plan for change management and user support.

Fox’s team spends time evaluating the current use of standards and metadata in clinical systems and processes anddeveloping a sustainable approach to metadata driven data curation and shared information models. They pilot and demonstrate feasibility of tools and approaches centered on FAIR data principles and processes. They drive collaboration across R&D to establish and evangelize the use of semantic tools in metadata management.

Following is the conversation we had with Julia Fox.

Delivering quality data is the core of clinical trial and data management in drug research. The key for data quality is to adopt common standards, stick to the standard practices and processes, and use industry strength technology to reduce human errors. Adopting a standard not only increases interoperability and efficiency but also provides a foundation for data integration and increases degree of code reusability.Your views on this.

Providing access to these common standards to systems and directly to users is the first step in promoting compliance in the use of enterprise standards. A centrally governed and managed source of Standards and Metadata values is critical to achieving harmonized and consistent use of standards and enabling semantically driven data models which dramatically increase data quality. Rich, meaningful and programmatically enabled semantic modelling drives interoperability, and can alleviate the burden of manual processes and curation to achieve high quality data and efficient data systems. Key tools in supporting semantic harmonization are entity-attribute-value models and ontologies to manage metadata objects, facilitating object-oriented modelling and data processing.

  ​Ontologies are the single most critical component in semantic modelling to promote data quality, as this is the nexus of object harmonizatio 

Ontologies are the single most critical component in semantic modelling to promote data quality, as this is the nexus of object harmonization. Comprehensive and complete annotation of values in ontology includes known synonyms, mappings to reference and internal system IDs, relationships, and contextualization. Instantiating these in a programmatically available format allows for easy use via APIs and microservices, enabling point-of-service, fit-for-purpose use of sematic models and harmonized values.

What is the importance of system integration which is an engineering process of bringing together the component subsystems into one system to deliver the overarching functionality and ensuring that the subsystems follow standards and function as a whole?

In my opinion direct integration between systems has become less necessary and more cumbersome than a modern ‘micro services’ approach. Direct integrations are labour intensive and require maintenance and oversight to achieve synchronicity and promote updates and changes. API connectors and web-base tools used to consolidate across systems is a more facile approach, alleviates some of the overhead in set up and maintenance, whilst allowing for synchronization and automation. Driving semantic interoperability from a centrally managed source which can ‘inform’ downstream systems within a data ecosystem allows conceptual integration without depending upon direct system integration.

What would be your piece of advice for your fellow peers and leaders?

Understand and integrate semantic management in Data management strategies to promote harmonization, synchronization, and usability of Data in systems. Metadata driven processes promote automation, machine learning and traverse-ability within and between datasets. All Data systems require semantic configuration, doing so from a centrally managed metadata source integrates information on the Data level rather than the system level, and enables faster analytics and more meaningful insights. When analysts can iterate quickly on models and evaluate parameters and data points in real time, the ability to refine and improve analytics for insights is accelerated, and the time to discovery is decreased.