Standardized Data Structures in Rare Diseases: CDISC User Guides for Duchenne Muscular Dystrophy and Huntington’s Disease

Abstract

Interest in drug development for rare diseases has expanded dramatically since the Orphan Drug Act was passed in 1983, with 40% of new drug approvals in 2019 targeting orphan indications. However, limited quantitative understanding of natural history and disease progression hinders progress and increases the risks associated with rare disease drug development. Use of international data standards can assist in data harmonization and enable data exchange, integration into larger datasets, and a quantitative understanding of disease natural history. The US Food and Drug Administration (FDA) requires the use of Clinical Data Interchange Consortium (CDISC) Standards in new drug submissions to help the agency efficiently and effectively receive, process, review, and archive submissions, as well as to help integrate data to answer research questions. Such databases have been at the core of biomarker qualification efforts and fit-for-purpose models endorsed by the regulators. We describe the development of CDISC therapeutic area user guides for Duchenne muscular dystrophy and Huntington's disease through Critical Path Institute consortia. These guides describe formalized data structures and controlled terminology to map and integrate data from different sources. This will result in increased standardization of data collection and allow integration and comparison of data from multiple studies. Integration of multiple data sets enables a quantitative understanding of disease progression, which can help overcome common challenges in clinical trial design in these and other rare diseases. Ultimately, clinical data standardization will lead to a faster path to regulatory approval of urgently needed new therapies for patients.

Scoping review of the recommendations and guidance for improving the quality of rare disease registries

“…However, the main challenge around common data elements is reaching a consensus regarding the choice, organization, and definition of the various elements [ 25 , 70 , 71 ]. Beyond simply determining the composition of the common data elements, other challenges include data coding standards (e.g., integer, float, string, date, derived data, and file names) [ 13 , 72 , 73 ], standardized data constructs, vocabulary and terminology [ 28 , 33 , 37 , 65 , 71 , 74 ], defined variable interpretation to avoid inconsistency (e.g., sex – genotypic sex or declared sex) [ 18 , 75 ] and ontology harmonization to facilitate convergence from different terms or languages [ 56 , 65 , 76 , 77 ]. The latter necessitates consistent agreed-upon disease classification standards [ 23 , 50 , 77 ].…”

“… Common data elements • Minimum or core group of data elements which can be expanded to meet the informational needs or objectives of the registry [ 37 , 41 , 50 , 53 , 56 , 69 , 76 ]. • Process describing how to define and organize the core data elements [ 13 , 25 , 70 , 71 , 74 , 75 , 77 ]. Data dictionary • Data dictionary provides clear instructions for data entry [ 17 , 20 , 25 ].…”

Clinical Data Interchange Standards in Clinical Trials on Alzheimer’s Disease

“…Using standardized clinical data models, a previous study could build integrated databases for the generation of drug development tools, such as polycystic kidney disease. 49 They used CDISC SDTM and the polycystic kidney disease-TAUG to map data from several academic registries and natural studies, and they developed a joint biomarker dynamics and disease progression model to demonstrate the relationship between total kidney volume and loss of kidney function by using integrated datasets. 49,50 By using structured and standardized data models and outcome measures on AD trials, the integration of several clinical trials on AD will be effective.…”

Fine-tuning Large Language Models for Rare Disease Concept Normalization

“…The presence of heterogeneous and unstandardized clinical data combined with insufficient comprehensive rare disease data only hampers and reduces the efficiency of medical research, which consequently may compromise the quality and reliability of clinical findings [5, 6, 7, 8, 9]. The adoption of a standardized vocabulary holds the promise of simplifying clinical data significantly, allowing researchers to easily compare and analyze data across multiple medical settings and databases, accelerating medical research [10, 11, 12].…”