Abstract
Background/Aims This presentation will focus on challenges faced by new HMO Research Network (HMORN) sites and their experiences with the Virtual Data Warehouse (VDW) tables for version 3 specifications. The step-by-step procedure developed by Scott and White will be showcased to provide a head start for new incoming sites. Sharing lessons learned during our one year in the HMORN VDW, our objective is to provide best practices regarding resources, time allocation, priorities, schedules, financing, backups, and possible available alternatives.
Methods The comprehensive procedure followed by Scott & White to build VDW V3 tables:
-
Setting up your SAS Environment
-
Establishing an architecture and components
-
Analyzing and forecasting server specifications;
-
-
Identifying and understanding the multiple data sorces embedded in your organization
-
Claims data
-
EMR
-
Lab
-
Pharmacy
-
Other;
-
-
Detailed variable mapping of sources to VDW specifications;
-
Assessment of source data availability;
-
Getting your site’s first feed from the source and following through with quality checks;
-
Methods for building transformation logic for the table;
-
Choosing the methodology to approach table-building.
Results We will showcase Scott & White metrics for building a VDW starting with version 3 and best practices used. We improved the validity and completeness of the data through person-to-person interactions with claims personnel who often were disinterested in the needs of research. Having strong advocates who think of research as a part of the institutional mission was helpful. The VDW data team included a sole programmer for the first 6 months along with the Site Data Manager; adding a second programmer improved the rate of progress.
Conclusions A structured, step-by-step, approach is recommended for new incoming sites joining the HMORN. This approach will optimize efficiency and effectiveness. The more information and guidance the existing members can provide, the better the achieved results will be. Sharing experiences will guide new members to a more pro-active and efficient start and ensure the success of seamless data-sharing.
