Implemention Technology
Systems Integrator
  Dataspace Incorporated
Business Intelligence
  Oracle Discoverer
Data Extraction
  Ardent DataSage
Data Modeling
  Computer Associates ERwin and
    Model Mart
Warehouse Engine
  Oracle
Hardware
  Compaq Alpha 8400 Server

The Detroit Medical Center (DMC), headquartered in Detroit, Michigan, is one of the nation's most prestigious urban healthcare centers. The DMC is the largest healthcare provider in southeast Michigan and operates eight hospitals including Children's Hospital of Michigan, Detroit Receiving Hospital, Grace, Harper, Huron Valley-Sinai, the Rehabilitation Institute of Michigan as well as two nursing centers and more than 100 outpatient facilities throughout southeast Michigan.

The DMC, as with many other healthcare organizations, is experiencing dramatic change. Declines in patient volume and revenues are driving a need to manage and tightly control operating costs with healthcare organizations. In order to survive in the dynamic healthcare environment, DMC management realized it was crucial to be able to access viable and timely cost data in order to make key management decisions. At the core of this strategy is DMC's decision to retain Dataspace, a data warehouse consultancy, in order to harness the power of advanced computer technology through the design of an organization-wide data warehouse.

The Procedure

As the data warehouse project was launched at DMC, so was the development of a Clinical Information System (CIS). The goal of using the same platform for both the data warehouse and the CIS system was to provide DMC the resource to leverage their purchasing volume of new hardware and software for better pricing. CIS and the data warehouse were the first ventures into client/server computing for the DMC. This made platform consistency a requirement to simplify maintenance for both the hardware and network. Hardware purchase decisions were primarily driven by the needs for the CIS system.

The DMC data warehouse transformation server, a NetFRAME NT Server by MicronElectronics, receives data from an RS/6000 that acts as the interface server for all system interfaces. Common system data, such as standardized patient room and bed locations and physician data is received from the mainframe via FTP. The data is cleansed and validated on the basis of DMC business rules. Then it is integrated and transformed before being loaded in the data warehouse database which resides on a Compaq Alpha 8400 server. Systems communication occurs across a mixture of ATM and ethernet, using both Novell and NT networking software.

Two hundred programs written by DMC staff capture a "super record" at the transaction level from SMS' Invision patient management system. Medical Computing's Link Receiver was customized to receive admitting, discharge and transfer transactions in the form of HL7 segments on the NT server.

From there, Ardent's DataStage was used to cleanse and restructure the data prior to loading it onto an Oracle database on the Alpha server. DataStage allowed the DMC to view its data in a common framework and also provided the flexibility to address the complex business rules inherent in a healthcare environment.

Oracle Discoverer was selected for reporting and analysis because it satisfies many reporting needs and allows DMC to leverage Oracle support across products. Generalized templates were developed and shared across the user community to simplify use. As both the warehouse and the user community's needs mature, a more robust OLAP tool ­ with better graphics ­ may be needed.

ERwin and ModelMart from Platinum (now Computer Associates) manage data modeling and database design. Both entity-relationship diagrams (ERDs) and dimensional models (star schemas) were developed using this tool; and, because the models could be shared, the ERwin models became the "source of truth" for what the final database needed to look like. Because all database changes were coordinated through ERwin, the model and the database were always synchronized. If there was a question about validity of the database structure, a referral to the ERwin model provided the answer. Consequently, this simplified the maintenance of changes in the database.

Learning the software tools was relatively simple for people who were new to the products. But, we experienced quite a few bumps in upgrading to "latest releases." As the source system Invision changed, the interface programs also had to change. Rigorous testing was done for each software upgrade, and several data issues were "caught" by the warehouse even though they had not been recognized in upstream systems.

A Healthy Outcome

DMC had many different systems in use at its various facilities that produced similar management reports but often with conflicting results. The systems were not integrated and they were difficult and costly to maintain. With Y2K just around the corner, it was imperative to find a data warehouse solution. We didn't quantify solutions, but we knew that the current reporting system had to be replaced. It was just costing us too much time and money.

While the initial goal of the data warehouse was to replace a patient management reporting system that was not Y2K compliant, the vision and goals for the warehouse have grown significantly. Prior to the warehouse, DMC never had the ability to look at patients and registrations across their operating units. Now, these can be calculated more accurately in the warehouse with a perspective of all DMC units as a single enterprise.

The mere visibility of warehouse information is causing a groundswell of demand for even more information ­ from medical record data to purchasing details. Plans for the integration of both of these sources of data are underway. The next warehouse release will tie patient accounting information to the patient visit. It's exciting to see how excited people become once they realize how positively the data warehouse will affect their work. They realize how much more effective the organization will be and how much faster and accurately decisions can be made.

Case Review

Timing is everything when it comes to launching technology initiatives. The DMC, already dealing with Y2K issues, chose to initiate both the CIS and the data warehouse project simultaneously. Consequently, the warehouse team had to compete for resources as well as for executive attention at critical project junctures. Luckily, the team had a strong project manager who was both technically savvy and politically skilled to guide the team through these crucial periods. This leadership and critical thinking allowed the project to move ahead as scheduled and not become mired in the corporate backwater.

Hardware and software selection should be done carefully so that your own business needs are addressed. Build in time to test drive the tools, either with evaluation copies or through computer-based tutorials. With reporting tools, make sure someone tests both the administrative side as well as the reporting side before making a decision. This extra effort on our part resulted in tool selections that are effective, easy to use and meet requirements.

Future Prognosis

The first implemented phases of the data warehouse have generated a strong demand from a variety of users throughout the DMC system. Access to meaningful information is the key to success for any organization in today's business environment. Our new warehouse is making this possible. The data warehouse is already a key decision platform for DMC personnel and as the user base increases, it will become integral to decision making at all levels of the organization.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access