The digitization of healthcare has made a flood of real world data (RWD) available for collection and analysis. RWD comes from electronic health records, medical claims and billing systems, patient-generated data, product and disease registries, and many other sources. Once consolidated and analyzed, RWD can be used as real world evidence (RWE) in various ways to improve patient outcomes and accelerate the development of medical products.
Capturing and effectively utilizing RWE is important across the healthcare value chain— from life science companies and providers to payers and regulators—but due to the variety and range of data sources, RWE data has been traditionally difficult to harmonize and reconcile for practical application.
Until now, it has been challenging to standardize and store operational data models for RWE. In addition to the lack of consensus around standard data models for RWE among the healthcare community, harmonizing real world data is particularly difficult because RWD sources capture observational content and differ in data quality, purpose, and design. In practice, RWD sources are often manually queried and analyzed with ad-hoc processes for specific insights.
An end-to-end approach combining human expertise and machine learning
The Tamr RWE Integration Solution provides an end-to-end approach to integrate and apply RWD into various life science functions. Through the combination of human expertise and machine learning, Tamr provides a robust platform to sustain ongoing integration of real world data sources into a unified data model, such as the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM).
With Tamr, companies can manage all varieties of data schema models throughout the data integration process and leverage built-in starting model templates, such as OMOP CDM, to accelerate their data integration initiatives. By leveraging human-guided machine learning, Tamr’s solution also provides automated mapping suggestions for on-boarding new datasets and correcting for potential schema drift.
The actual development of new data transformations is pretty straight-forward thanks to Tamr’s intuitive SQL-like syntax, and rapid data processing using a high-performance Spark engine that easily scales for billions of records. If a project has existing data harmonization specifications, existing mapping specs can be directly uploaded into the RWE integration solution. Throughout the data transformation process, Tamr tracks all data conversion transformations and user interactions to provide clear data lineage and audit trails.
In the end, the Tamr RWE Integration Solution offers numerous benefits. Among them, it:
Integrates messy, disparate data sources
Processes up to 30+ billion records a week
Improves R&D designs and analysis
Delivers competitive market insights
Automates and streamlines data analysis
Confirms drug safety and efficacy to regulators
Consolidates data conversion specs
Enables scientists and researchers to discover breakthroughs faster
Speed, scale, and sustainability is key for RWE integration
Many life science companies are limited by their ability to efficiently move data through their data infrastructure; most ETL solutions are lacking in the capability to regularly transform and process disparate data sources into unified views. Further hampering data integration efforts is that traditional ETL solutions may take weeks to months of set-up for new data sources and several days to weeks when processing large volumes of data varieties.
The Tamr solution, however, provides a wide collection of transformation templates and capabilities that can be reused across various data sources. Along with Tamr’s machine learning features to support data modeling, the set-up and maintenance for data integration becomes significantly faster. In addition, transformations are processed through Tamr’s innovative implementation of Spark that can be scaled to efficiently process RWD records at over 30+ billion records (or multiple terabytes) a week.
With the Tamr solution, life science companies can make a transformational shift in delivering results that were previously impossible. Tamr worked with one multinational life sciences company to integrate RWD from the Clinical Practice Research Datalink (CPRD), and data vendors, such as Truven Health Analytics and Optum, to construct a repository of RWE for mining analytic insights. Within weeks, Tamr was able to build a robust pipeline integrating 30 billion records a week from RWD sources into 100+ billion record outputs using the OMOP common data model.
The ability to capture and leverage RWE effectively is now within reach. As more and more companies adopt the application of RWE, the approach to healthcare will change and organizations will be better positioned to deliver transformational changes in the pursuit of delivering better treatments and outcomes for patients.