Pre-specified Events and Pre-specified Findings

Pre-specified Events

Collection of adverse event, clinical events, and medical history events can follow two approaches:

  1. Were there any events?
    1. If yes, what were the events?
  2. Did event X occur?
    1. If yes, record the details of the event(s)

The first approach is sometimes called “spontaneous” reporting. The answer to question 1 is in a CDASH --YN variable but is not recorded in SDTM. The results of the follow-up question 1a are represented in records which include the name of the event in TERM, and information about timing and other characteristics in other variables.

The second approach records pre-specified events. The fact that the question is of this type is represented by the inclusion of the variable PRESP with a value of “Y”. The name of the event ("X") is in TERM and the response to question 2 is in OCCUR.  If OCCUR = "Y", then the timing and other characteristics of X (the details in 2a) are recorded in other variables in the same record.

In summary:

  1. In the “spontaneous” approach, TERM is an answer to question 1a, with further information about the event in other variables.
  2. In the “pre-specified” approach, TERM is part of the question, the answer to question 2 is in OCCUR and further information about the event is in other variables.


Pre-specified Findings

For examinations that look for abnormalities, we have a similar situation.  Questions can be asked as

  1. Were there any abnormalities?
    1. If yes, what were the abnormalities?
  2. Did abnormality X occur?

In the first approach, sometimes described as an “open-ended question”, the answer may or may not be represented in an SDTM record. An example where it would be represented in SDTM is for ECGs, where a record for test "Interpretation" would have a value of "NORMAL" or "ABNORMAL.”

The results of follow-up question 1a are represented with a broad TEST value (e.g., "Microscopic Examination", "Physical Examination", "Sediment Examination" or, for ECGs, tests such "Axis and Voltage" or "Sinus Node Rhythms and Arrhythmias"). The particular abnormality is in the result variables. In the second approach, the “pre-specified question” approach, abnormality X is part of the test and the result is "Y" or "N" (or "PRESENT" or "ABSENT" or "INDETERMINATE", etc.).

In summary:

  1. An event always goes in TERM, whether or not it is part of the "test", and there is an extra variable (OCCUR) to hold the answer to the pre-specified question.
  2. For findings about abnormalities, the abnormality goes in result variables in the first approach, but test variables in the second approach. The result variables show have an abnormality with the first approach, but a Yes/No with the second approach.