VistA Data Model

VistA Data Model » VISN 20 File Directory

VISN 20 File Directory

FileMan is VistA’s proprietary database. It holds not only clinical data but also nearly all the configurations, audit logs and financial information of the system. Having all this data in one location is a boon for VistA system analysis and the centrality of FileMan effectively turns VistA analysis into FileMan content analysis.

In FileMan, data is gathered into “files” which are structured by “data dictionaries” or schemas.

The following reports on a VISN 20 File Directory built from the data dictionaries of 8 VistAs: Anchorage [463], Boise [531], Portland [648], Roseburg [653], Puget Sound [663], Spokane [668], Walla Walla [687], White City [692]. Together the systems hold 6,776 distinct files containing 7,731,166,421 records.

Files can be grouped into those in every VistA (“all”), those in more than one but not all (“some”) and those only in one (“exclusive”). The files in every VistA represent the core of a VA VistA, that portion of FileMan common to production systems …

visn20AllFileTypesIsHasBH

Files can also be grouped on whether they are “empty” or have only one record (“singletons”) whereever they appear. The empties usually represent retired parts of VistA and can be put aside …

visn20FilesTypesInUseIsHasBH

Putting aside the empty files puts focus on the active parts of the VISN 20 VistAs. See the contrast between common or core files and the proportion of records in those files - while only 2,792 (47.48%) files are common to all systems, those common files contain 95.73% of the records in these systems. In other words, the core of a VistA FileMan holds most of the records …

visn20FileNRecordCountsByFileTypeCategoryBH

Put side by side, you see that systems big and small have very similar numbers of files but that larger systems like Portland (“POR”) have proportionately more exclusive files …

visn20SideBySideCommonCountsNoEmptiesCategoryBSV

Record numbers, side by side tell a different story. The biggest system is Puget - not as file count indicates, Portland - and Puget is responsible for most of the populated exclusive files. In other words, though all systems have some exclusive files, they are of most (only of?) importance in Puget …

visn20SideBySideByCommonRecordCountsCategoryBSV

VistA Clone Details

Data from 8 VistAs was parsed. The VistAs are copies of production VistAs cut on the dates given below …

# VistA Station # Clone Cut Date
1 Anchorage 463 2021-01-06
2 Boise 531 2021-05-08
3 Portland 648 2021-02-12
4 Roseburg 653 2021-05-09
5 Puget Sound 663 2021-02-16
6 Spokane 668 2021-05-10
7 Walla Walla 687 2021-04-18
8 White City 692 2021-05-08