Translational or “bench to bedside” research aims to bring laboratory research and findings to patients in the form of new treatment options. Patients are in need of precise treatments tailored to their unique OMICS profiles. The latest research domain involves computationally analyzing genomic and extracted clinical data from Electronic Medical Records (EMRs) to determine how to prevent and treat diseases in individual patients. However, obtaining data ready for analysis from different hospitals and EMRs proves to be a daunting task, one that requires expert knowledge and manual handling. Interoperability of Healthcare data from different sources remains an industry challenge.
Information for Integrating Biology and the Bedside (i2b2) was first launched in 2004 by National Institute of Health (NIH) to expedite the sharing and reusing of clinical patient data from individual institutions. It provides an ontology based object-oriented, highly flexible, star-shaped data schema consisting of a central fact table surrounded by multiple dimension tables containing further attributes. Several ontologies can be used, which is ideal for collecting data from a variety of different sources. Data can be entered directly into the database, and i2b2 also has the ability to run analyses and reports and other visualizations. Now in version 1.7.10 (May 24, 2018), i2b2 has served as an excellent tool for translational research and recently merged with the transmart Foundation. The Initiative has been adding more and more analysis and visualization tools to their repository over the years.
A subsequently launched data model is the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM), created by Observational Health Data Sciences and Informatics (OHDSI) in 2009. OMOP CDM allows for the systematic analysis of disparate databases through the conversion of data to a common format and representation – this means common terminologies, vocabularies, coding schemes, etc. Therefore, a variety of different ontologies can be supported. The person-centric relational data model consists of domain tables containing concepts in an entity-attribute-value structure. Standard analytic routines are then used based on the common format: OHDSI also has developed ATLAS, a comprehensive set of tools for such analysis and visualization of data. Though the Extract Transform Load (ETL) process of data can be cumbersome, there are several additional tools to facilitate the process:
- White Rabbit: inventorize your source data (what tables and variables are present)
- Rabbit in a hat: document/visualize the mapping of the source data to the data model
- Usagi: map concepts used in the source data to the ontologies (vocabularies), for instance for medication, treatments etc
- Achilles: evaluate the upload of the data and clean the data
The beauty of OMOP CDM is its ability to share analyses and results with other users without sharing the actual data itself.
Ultimately, data models such as i2b2 and OMOP CDM are crucial in translational research and precision medicine. Without such tools, analysis of patient data to develop new methods of prevention and treatment for diseases could not progress as they currently are.
www.thehyve.nl provides an excellent review of several new translational biomedical informatics softwares such as RADAR-base (scientific research through wearable data collection and analysis devices), CBioPortal (visualization and analysis of large-scale cancer genomic datasets), and MatchMiner (algorithmic matching of patient profiles and clinical trial criteria).