3 Replies Latest reply on Dec 20, 2016 11:36 AM by Shinichiro Murakami

    Dimension representing "contains" rather than "is"

    Jon Seymour

      Vexing question from a new user. I have a field that contains multiple values. (Specifically, it is ICD-9 medical codes for diagnoses.) The values are delimited by ^. Each record contains from 0-45 diagnoses. More than half have only one code, and there is a distribution similar to what one would expect, with fewer containing two, etc.

       

      What I am trying to do is graph the most common diagnoses. I started by creating a split, which gave me an index, if you will, of the ICD-9 codes. I then placed that into the rows dimension. A count of the records by claim number is in the columns dimension.

       

      What I see in the graph is not a count of all the records containing each ICD-9 code. Instead I am seeing a count of all the records containing ONLY that ICD-9 code, i.e., the exact match.

       

      So, I cannot figure out how to get the rows to represent "contain" rather than "is".

       

      Using a filter, I can get counts of records containing a single ICD-9 code by using the "CONTAINS <code>" function...this works correctly, giving me the full count including contains and exact..but this only works one at a time.

       

      Any ideas on how to achieve my goal of counting up and graphing the most common diagnoses by frequency?

       

      Thanks!