Automat - Central Registry Unduplicate Procedures

Mark Combs
October 2003

From time to time we use automatch software to run an unduplicate match against the central registry database. This allows us to locate duplicate patients that may have different spellings of names and to then consolidate that information into one working record.

Here are the steps to find duplicates in the central registry:

When central registry is not being used run data analysis using a query that will select all the records in the database.
Run data exchange with the format file "dupcheck.fmt". This file can be found in /export/user/mcombs/automat/kcr_undup. The output file should be directed to "dupcheck.out".
Secure FTP the file from the directory you ran from on ALPHA to J:\automat\kcr_undup on PHOENIX. NOTE: you may want to rename the old dupcheck.out file on PHOENIX instead of overwriting it.
From J:\automat\kcr_undup run "match.bat". This will automatically run the first level of matching.
Run "clview dupcheck 1", this will start the clerical review for the first pass. Review possible matches and flag them appropriately.
Run the second pass of matching "mtch dupcheck 2 1 > dupmtch2.txt".
Run "clview dupcheck 2", this will start the clerical review for the second pass. Review possible matches and flag them appropriately.
Run "repgen dupcheck" to create reports "clerrpt.txt" contains the patients that need to be reviewed further, "duprpt.txt" contains the patients flagged as duplicates.
Pull these reports into textpad, copy just the lines you need and paste into a blank excel sheet so you can have more control for printing.