Based on a thorough review of current practices and data used, Redshift provided a simple “open and flexible” container approach to support Pfizer’s unique data ontology. This allowed for the agile development of common data categories and areas by subject matter experts and specialized views for the data community. The first step Redshift employed to address a lack of common data standards, was to define common naming conventions. These then became the root level from which all data would then be defined. Redshift worked with Pfizer to develop and define common data categories and areas so that unstructured data could be harmonized and reside in usable families.
As an example, to avoid duplication and errors, a single name was identified for each compound root and synonyms were linked with the appropriate name. Compound research documentation was then transferred from disparate locations into the Redshift Profiler, and assessment outcome data (metadata) was extracted from gathered research. Data was formatted following standard federation rules where available, as is the case for chemistry. In cases such as worker safety, where standard formats were not available, rules and formats were created.
Intricate logic models were programmed to ensure the data would be analyzed, merged and compared accurately and insightfully. The user interface and the reporting options were programmed to ensure usability and accurate insight into compound data. Finally, privacy and security requirements were programmed to ensure appropriate access to data.
A Centralized Database of Compound Data
Today, the Redshift Profiler provides the foundation for the Pfizer EHS groups to share valuable compound assessment data:
- When toxicologists acquire component or raw material data in any format, they enter them into the Redshift Profiler with a consistent data structure. They also attach the original source document so it is available for any scientist within EHS and PDM to verify any details.
- A central database contains links to all available research on a compound, and security levels for each component assure only appropriate users access the data. Toxicologists can also share data with upstream and downstream systems.
- Extracted outcome data are formatted to provide high level views, analyses, comparisons and reports.
- Scientists can adjust parameters in the original data at any time they deem necessary. As more scientists evaluate the data, multiple opinions and additional studies are added to the shared knowledge pool increasing its value. The data lifecycle tracks how the data were generated over time.
Reduced Exposure to Worker and Environmental Risk
By capturing and sharing compound knowledge across the organization, Pfizer has reduced its data inaccuracy and their ability to develop hazard and risk profiles for their customers. The new system has:
- Reduced the time it takes EHS experts to complete compound assessments:
- Now that Pfizer scientists have access to all existing research about a compound gathered or conducted by the organization, they have a better understanding of the hazard associated with a compound and any testing gaps that need to be addressed.
- Real-time access to all research data available has minimized duplication of compound evaluations and facilitated the repurposing of existing data.
- Improved the reliability and access to data while reducing the labor associated at with treasure hunting for data:
- Data are more accurate as subject matter experts throughout the organization with accounts can add specific endpoint data, and vary the weight of a data point in an assessment depending on their level of data confidence gleaned from the system.
- Worker safety data are available real time and viewed throughout the organization by any person with an authenticated Pfizer domain login.
Share with Third Parties
In the interest of increasing worker safety and reducing environmental risk through assessment accuracy, Pfizer has plans to continue to expand the capabilities of the Redshift Profiler. Next on the list is an XML access/ feed for third parties to view and upload data. The first step will be to allow external organizations access to designated Pfizer information in the Redshift Profiler, and the option to enter additional compound research and opinions. These third parties could be customers, agencies like the EPA, and also competitors with whom Pfizer could share non-proprietary research. Users of data, as for example, risk assessors, can access information about the hazard and risk for a particular compound via XML. If there an assessment is not available, they will know immediately and can request it directly.