Comparing Dictionaries

Overview
One of the main tools in maintaining consistency and proper change management over your dictionaries is the ability to perform detailed comparisons.  This can be applied both to thesaurus or controlled coded terminology dictionaries.  The differences among the dictionaries shown through comparisons can reveal discrepancies.  Reconciling the discrepancies help you maintain optimal accuracy.  The comparison tools can be accessed from the main screen as shown here:

The main steps in performing comparison for thesaurus dictionaries include:

The main steps in performing comparison for coded terminology dictionaries include:

Selecting Thesaurus Dictionaries
The first step in performing comparison is to select the thesaurus dictionaries.  ThesQA manages two types of dictionaries including Thesaurus and Coded Terminologies.  In this case, select the dictionary type of "Thesaurus Dictionary".

There are two dictionaries that will be compared. They are identified by their names as in "Dictionary Name (1)" and "Dictionary Name (2)".   The name is also used as a unique identifier so you can use this as a way of selecting the dictionary through the pull down menu.  The other information pertaining to the dictionary is displayed but is grayed out and not editable.  These include:

  • Path - Physical path to where the dictionary information is stored
  • Data - SAS dataset which stores the dictionary information
  • Data Label - The label of the data documenting the description of the dictionary
  • Keys - These are three unique key fields that are used to merge and sort the data during comparisons

There are different categories of thesaurus dictionaries including items like WHO Drug or MedDRA.  The selection of the first dictionary will alter the selection of the second dictionary since you will only be able to select dictionaries from the same category.

You can also choose to preview the data by clicking on the preview button next to the data entry.  This will display the first 100 observations of the contents of the dictionary itself.   

Performing Thesaurus Dictionary Comparisons
Once you have selected the two dictionaries from the COMPARE screen, you can perform the comparisons by clicking on the OK button.  This will perform a sort and merge between the data of the dictionaries to find the differences.  A report is then generated documenting its findings with the following key areas:

  • Compared Dictionary Details - Document the attributes of two dictionaries including information such as data set and keys fields.
  • Mismatched Observations - Summarize all the rows in the dictionary that has matching keys between the two dictionaries.
  • Detailed Summary - Details all the rows that are matched and unmatched between the two dictionaries

Saving Thesaurus Dictionary Macro Code
The same selections that is made in the graphical user interface can be expressed in macro named %tcompare.   You can select this through the "Save Code..." button.  The sample code that is generated would look like:

libname Thesqa "C:\Program Files\MetaXceed\ThesQA\data";

%tcompare(thes1 = Who Project Specific,
                 thes2 = Who X Test,
                 dictype = thesaurus);

 
This same program can then be executed in batch mode to perform the same dictionary comparisons.

Reviewing Thesaurus Dictionary History
All the comparisons between thesaurus dictionaries are retained in a complete audit trail.  You can view this history by clicking on the "History" button.  

 

This view presents all the comparisons that have been done in chronological order.  This includes all the attributes of the comparisons including who has applied it and at what time.  A search can also be applied through the search button from the compare history view button.

The search criteria is then applied to all the fields in the history to derive a report capturing all of its findings. 

Selecting Coded Terminology Dictionaries
The first step in performing comparison is to select the coded terminology dictionaries.  ThesQA manages two types of dictionaries including Thesaurus and Coded Terminologies.  In this case, select the dictionary type of "Coded Terminology". 

There are two dictionaries that will be compared. They are identified by their names as in "Dictionary Name (1)" and "Dictionary Name (2)".   The name is also used as a unique identifier so you can use this as a way of selecting the dictionary through the pull down menu.  The other information pertaining to the dictionary.  These attributes include:

  • Coded Name - A unique identifier assigned to the set of coded terms
  • Coded Value - A specific coded term that is associated to a coded name
  • Decoded Value - A specific decoded value associated with the coded value

The selection of the dictionary name will refresh the list of attributes pertaining to the coded terminologies.  Besides selecting the dictionary name, you must also select one specific coded name for each dictionary to be compared.  This is accomplished by simply clicking on one of the names.

Performing Coded Terminology Comparisons
Once you have selected the coded name for both source and destination dictionaries, you can begin performing the comparison by selecting on the OK button.  A report will then be generated which includes the following information:

  • Matching Terms in Both Dictionaries - This report reveals all the codes and decodes that have matches between the coded and decoded terms.
  • Mismatch Terms found in Source but not found in Destination - This report displays all terms in the source that were not found in the destination
  • Mismatch Terms found in Destination but not in Source - This report displays all terms in the destination that were not found in the source

Saving Coded Terminology Macro Code
The same selections that is made in the graphical user interface can be expressed in macro named %tcompare.   You can select this through the "Save Code..." button.  The sample code that is generated would look like:

libname Thesqa "C:\mypath\data";

%tcompare(thes1 = CDISC 3.1.1 Controlled Terminology,
                thes2 = Sample Coded Terms,
                dictype = coded terms);

This same program can then be executed in batch mode to perform the same dictionary comparisons.

Reviewing Coded Terminology History
An audit trail of the comparisons between dictionaries is kept in the history.  This history can be reviewed and searched through the "History" button.  A history view looks like the following:

An audit trail captures all the terms that have been applied including when and who has applied it.  This information can be searched by any field.  The search criteria in which you specified will be applied to all the fields captured in the audit history.

 
      ThesQA - Manage Controll Terminology, Meta-Xceed Inc.© 2009
Bookmark and Share