Learning Objectives for this Presentation:
By the end of the presentation participants will be able to help practices help themselves improve data quality by identifying patterns of errors, and providing data improvement support.

Background:
The Vermont Immunization Registry began as a user-interface-only system. To date, all data in the IMR was entered by individual users at provider practices or WIC clinics.

Objectives:
To improve the quality of data in the Immunization Registry.

Methods:
With the release of our data warehouse and the acquisition of data quality software, we began running monthly quality checks on registry data. Our initial data checks focused on duplicate records, and apparent over-immunization. Many of the errors we identified were idiosyncratic to specific practices. We developed a procedure to address these errors on a by-practice basis by incorporating a data improvement element to our “care calls,” or practice visits.

Results:
Effective data quality improvement involved three steps:
• Identifying patterns of data entry errors.
• Assisting practices with data cleanup.
• Working with practices to change patterns.
The Registry team found that approaching practices with a prepared list of questions and working with the users to discover why errors had occurred was more effective than simply indicating the errors and leaving them to the practice to resolve on their own. The practice took ownership of maintaining registry data quality, but the Registry team took responsibility for providing data quality oversight and training users to use the application as intended. While providing the correct data was the practice's responsibility, the registry team took responsibility for identifying patterns of errors and guiding the users toward error-free registry use.


Conclusions:
This approach not only helped correct many of the problems with registry data but also provided encouragement to users since in many cases, entering data correctly actually meant less work for users.