Anticlue

Data Loading a Warehouse

One of the problems with interfacing hospital data from system to system and all over the place is data integrity and data ownership. Ensuring synchronization between all of the systems is several peoples full time job, especially when the data is being evaluated for reporting, quality initiatives, and patient evaluation. I’ve always thought a data warehouse including the clinical, labs, pathology, orders, registration, billing, abstracting, case management, denials, and other system data stores is a great place to centralize the data. The structure of a data warehouse is dimensional with fact tables, so data can be analyzed over dimension compared with facts. I think it would be an easier mechanism to utilize for reporting, and allow the users to create reports adhoc.

One of the obstacles has always been the data load, finding a tool that can accurately extract the data from the operational data stores, and place the transformed data into the data warehouse. Cold Fusion could be that tool, especially with the plethora of JDBC drivers. CF is excellent at manipulating data, and extracting data, and CF can connect to multiple data sources. Then load could run off of stored procedure in an ORACLE or DB2 db, using the files from CF extracted from the system, or a temporary SQL Server DB loading area. Does anyone use CF for the extraction process to a Data Warehouse? I think it would be a good tool to consider.