Life Sciences

d-Wise's View of Metadata Repositories (MDRs) – 2021 

Share:
Life Sciences

Metadata-RepositoriesContinuing our discussion on data warehouses and repositories in the clinical lifecycle, we turn now to the “new kid on the block”, the metadata repository (MDR). Metadata is a word that is thrown around without a real understanding of what it includes and how valuable it can be to your company’s process. It determines how pieces of information are related and provides tools for explaining/defining how data are related, so that it then becomes easier to turn this data into information about processes.

With the evolution of clinical data standards, metadata management has become all the more important and finding ways to better store, access, and manage metadata is regarded as a priority by most life sciences companies.

The Old – and Inefficient – Way

The original way of managing metadata was through spreadsheets, a process still used by many companies. But spreadsheets are not robust enough for this purpose and in general do not provide rigorous controls, which makes it difficult to leverage metadata to support and update standards, map processes and support relationships between data elements.

The Ideal Solution

 

On the other hand, an efficient, well implemented MDR has the potential to improve the submission process while reducing costs, improving data quality and compliance, improve business processes, and drive automation.By having the ability to populate the MDR with metadata from clinical studies, many of the manual processes of managing data become automated, streamlining processes.

Metadata-RepositoriesMetadata management is important for the development, implementation, maintenance and administration/governance of standards. Implementation of an MDR is not simply an implementation of the tool itself but also the institution of a governance structure to administrate standards development, the development of processes that formally require and manage usage of those standards, the adaptation of any tools/systems important to the business functions to utilize and propagate those standards, and re-education of both management and operational teams.

What does it mean?

So this all sounds wonderful and every company should head down to your local office supplies store, swipe a credit card, and purchase an MDR. Unfortunately products that support the concept of an MDR to manage your metadata and support your processes have only recently been released to the market and there is more maturation to be expected from all of the offered solutions. 

One of the problems has been that many existing tools on the market have tended to manage metadata as an afterthought, with the majority of tools being data-centric rather than focusing their key capabilities around storing, managing, and consuming the metadata.   Moreover, few tools have been robust enough to handle the relationship between various standards, such as ODM, SHARE, and HL7 standards.EntimICE - With this need and growing expectations, some MDR solutions are now emerging that are more truly metadata driven.

Key tools for MDR

Metadata-RepositoriesAmong these are entimo’s MDR offering entimICE, a modular solution platform covering clinical and pre-clinical phases of the drug development process. Among its features are a collaborative interface for different user groups, enhanced traceability and information visualization, and a statistical computing environment to enable programming and data management in a controlled and traceable manner. entimICE products enable a variety of tasks to enhance automation.These include: 

  • A data mapping environment that makes it possible to transform datasets to arbitrary target structures, including CDISC SDTM.

  • SDTM and ADaM checkers that inspect structure and controlled terminology run officially published and additional checks and provide detailed reports like result statistics.

  • A dedicated generator that streamlines the creation of Define.xml for SDTM, ADaM and other datasets 

  • And an ODM validator.

SOA Software has adopted a different approach to implementing standards while automating business processes: semantics management. The company’s Semantics Manager, which acts as a virtual hub within an enterprise, is designed to enable standards development organizations and Sponsors to develop, obtain and incorporate reviewer input, while managing and distributing standards.

While data standards are designed to improve the efficiency of the submission review process, implementation of standards can be difficult. SOA’s approach to overcoming such issues is semantic interoperability, wherein computer systems can exchange data with an understood or common meaning of the data without human intervention or interpretation. SOA’s Semantic Manager was recently selected by CDISC as the SHARE (Shared Health and Research Electronic Library) technology platform. SHARE will be a global electronic repository for developing, integrating and accessing CDISC metadata standards in electronic format.

SOA Software offers a common service metadata model and governance across its various products. Their Semantics Manager software platform enables organizations to develop, obtain and incorporate reviewer input, manage, and distribute standards more efficiently. It also helps organizations to implement standards faster and more effectively while automating and improving business processes. Essentially, it helps users to define, find, understand, use, and exchange clinical data and metadata. A base library including all CDISC Terminologies, CDASH 1.1, SDTM 1.2, BRIDG 3.1, and ISO 21090, gives life science organizations a jumpstart towards aligning internal information models with domain standards and regulatory requirements.  

As regulation and the number of standards increase, so does the time and effort it takes to develop, manage, and distribute these standards in a way that truly results in the benefits intended. By binding systems to robust common domain terminology produced and disseminated by a standards development organizations (i.e., Semantics Management), industry can more efficiently implement end-to-end data lifecycle processes.

Additionally, regulatory data submission guidelines, reporting mandates, internal business efficiency improvements and merger rationalization efforts are driving life science organizations to replace their existing manual, inefficient management and implementation processes with fully interactive metadata governance platforms. 

Accenture/Octagon Research is approaching clinical metadata management through its Quantum MDR solution that is designed to enable a global standards governance process to be built, deployed and maintained. The product addresses the challenges of managing the lifecycle of metadata across an R&D organization with a centralized data registry and a comprehensive development and maintenance environment that enables organizations to effectively create and maintain clinical metadata.

Metadata can be defined and tracked according to data domains, individual data elements and controlled terms. Quantum MDR comes configured with the CDISC SDTM domain model as a base configuration.  Quantum MDR is a scalable enterprise-level technology platform on which a global clinical data standards governance process can be built, deployed and maintained, all while ultimately providing the industry with the tools necessary to comply with mandated data standardization guidance.  

Let's Recap

While the solutions mentioned above are long-awaited, welcomed, admirable attempts by technology vendors to address the needs of industry, each advancement and feature, results in challenges with configuration, implementation and governance. 

For most established and emerging pharmaceutical, biotech and medical device companies, there is no mechanism for managing and communicating metadata standards. These organizations are challenged to develop a governance process, or global standard to convey to internal and external data stakeholders.

The absence of standards and the inherent process inefficiencies translate to compromised timelines, data integrity issues and an inability to scale, resulting in potential delays in the delivery of the product to patient population.Is the industry simply looking for a replacement for spreadsheets with better control? Are we considering all the other aspects of metadata independent of the clinical data such as process metadata? Can the industry take full advantage of the power of semantic models for metadata management?  These questions have yet to be answered and it will be interesting to see how this area evolves over the coming years.  

Life Science Experience - Case Study