How Do I Represent a Pre-specified Finding?

In an events or interventions domain, the variable --PRESP = "Y" can be used to indicate that the value in the topic variable (--TERM or --TRT) was pre-specified. However, --PRESP is not a variable that is allowed in findings domains.

Suppose a study collected the following data about findings from a chest x-ray:

  • Is effusion present? Yes/No
  • Are infiltrates present? Yes/No
  • Other findings, specify ____________

In a findings domain, the first two questions are treated as specific indicator tests with results of "Yes" or "No", while the "Other findings, specify" question is represented as an open-ended question with a free-form text response, as illustrated in this example. (Note that this abbreviated example table does not include all required and expected variables for the Respiratory System Findings [RE] domain)

re.xpt

    Row Description
    1-2: These records for pre-specified findings use indicator tests for specific findings. These two tests are in CDISC Controlled Terminology (CT); if data was collected for a finding not in CT, the implementer would create an appropriate test name and test code.
    3: This record uses the open-ended test for the respiratory domain from CDISC Controlled Terminology. This example includes only one "other" finding, but if multiple "other" findings were recorded, each would be represented by a record with this same test.
    Row STUDYID DOMAIN USUBJID RETESTCD RETEST REORRES
    1 ABC23 RE 101 EFFIND Effusion Indicator N
    2 ABC123 RE 101 INFLTIND Infiltrates Indicator Y
    3 ABC123 RE 101 REEXAM Respiratory System Examination Calcifications

    All three records in the example can be thought of as results of an open-ended examination of the chest x-ray in which all abnormalities are recorded. Some findings were collected in response to questions about pre-specified findings while others are collected as free text results in a representation of the data with certain drawbacks.

    • There is no one field where all findings are represented. 
    • For any pre-specified findings, an "indicator" test for that particular finding must be used.