Biomedical Engineering Reference
In-Depth Information
Data that are modified or transformed by the data dictionary are normally stored in a data mart or
data warehouse so that the transformed data are readily available for subsequent analysis without
investing time and diverting computational resources by repeatedly reformatting the data. In the
example in
Figure 2-9
, the relationship between microarray data and clinical data, such as activity at
a particular gene locus and overt aggression score, can be more easily computed because the data
can be sorted and compared by date of birth. That is, the more likely transformed data will be used
in analysis in the future, the more valuable the data warehouse and the data dictionary.
Archiving
Archiving, the central focus of the data life cycle, is concerned with making data available for future
use. Unlike a data repository, data mart, or data warehouse, which hold data that are frequently
accessed, an archive is a container for data that is infrequently accessed, with the focus more on
longevity than on access speed. In the archiving process—which can range from making a backup of
a local database on a CD-ROM or Zip® disk to creating a backup of an entire EMR system in a large
hospital—data are named, indexed, and filed in a way that facilitates identification later. While
