Databases Reference
In-Depth Information
FIGURE B.25
Local and global metadata.
common analytical data warehouse. But the data in the local data warehouse is far less voluminous
and far less robust than the data found in the larger collective common analytical warehouse serving
the entire medical community. Stated differently, the larger collective common analytical data ware-
house is made from the data found in hundreds of hospitals and hundreds of clinics. But the local data
warehouse has information specific to the local care provider only.
Figure B.26
shows that there are
different levels of data warehousing within the healthcare information factory community.
In the ideal case the local hospitals and clinics would create their own data warehouse, then pass
data from their data warehouse to the larger analytical data warehouse for all hospitals and clinics.
But building a local data warehouse is often beyond the means or ambitions of the local hospital or
clinic. Therefore, it often happens that the larger analytical warehouse is fed directly from raw data
found in the local hospital or clinic.
Data models and the healthcare information factory
The database design that is done for the different levels of data in the healthcare information factory
is initially shaped by the data model. Like all data models, the healthcare information factory data
model is an abstraction of the information found in the different data warehouses that constitute the
healthcare information factory. Some of the qualities of the data model are:
●
Data is organized at a low level of granularity.
●
Data is structured according to major subject areas.
●
Data is interrelated from one subject area to the next.
●
Different groupings of data have identifiers—keys that may be unique or nonunique.
The data model is divided into three distinct levels. Those levels are the high-level, the mid-level,
and the low-level data model.
Figure B.27
shows these three levels of modeling.
