Since the mid 1990’s Blue KC has used multiple systems built in house to assign an individual unique identifier. The individual unique identifier was stored with membership data in the Data Warehouse and used for reporting and extracts to external vendors for analysis. The legacy matching logic did not perform at a very high level. The process was designed to either match a member with an existing individual unique identifier or assign a new identifier. There was no manual review process for questionable matching results.
In 2012, we began a search for software that would improve our data quality from a company that would also be engaged in the migration process to their product.
ICW’s Master Patient Index (MPI) product was selected for several reasons
- The ICW staff’s knowledge of the product details, configuration, installation and migration.
- The ICW staff analyzed our data and tailored the product based on the data analysis.
- The MPI application was a web based application that used a SQL server database.
- The migration to the ICW product retained our existing MPI assignments.
- The MPI application included configurable scoring limits to move questionable matching scores to a manual mapping queue.
- The manual mapping interface was easy to use.
- The MPI application also allowed merging and splitting member records when MPI assignments were incorrect.
The migration of three million member records was done through the batch processing in the MPI application over a three months period. The MPI matching process worked extremely well and also showed where our legacy systems had incorrectly assigned a MPI. The biggest challenge with the migration was having to clear the manual mapping queue of 35,000 records. It took two dedicated people about two months to assign all the records. Since then the manual mapping queue processes around one percent of the total records process each day. The daily manual mapping queue takes less than 30 minutes to process. The MPI application successfully replaced our legacy system with no negative impact to the people and processes using the MPI assigned value.
The input data quality has a large impact on the MPI application matching process. Names, addresses, and even birthdates are inconsistently entered by the members for different enrollments. These changes usually push the matching work to the manual queue but on occasions, the matching score creates a new MPI for the input record.
The MPI application’s automatic matching, manual matching, and merging /splitting capability has improved our data quality. This has reduced the data correction to fewer than 20 instances in a year. Another improvement that is difficult to measure is the ability to merge an individual’s multiple MPI records created in the legacy system. This gives us a better view of the person over time.
The application has been solid in its performance requiring very little support since our 2012 deployment. The software upgrades have gone extremely well and were completed without disrupting the daily processing.
The MPI solution has enabled Blue KC to meet new demands for comprehensive, aggregated member/patient data in the health information exchange (HIE) environment as well as in achieving data stewardship goals within Member Services. As the healthcare rules change, it is critical to maintain high quality membership information for accurate analysis and member engagement for the quality of their health.