Redcap notes from Doug

REDCap insights from meeting with NCANDA and ABCD Pitt DB admin

20230308: Questions for Doug/ABCD

  1. Especially useful calculated fields, automatic variables?
  2. how to set RecordID to be something besides 1,2,…
  3. change question on an instrument
    • VL: can hide old questions
  4. saving partially complete instruments AND redo
    • want to save whats already collected, but repeat to complete
    • replace? break big instruments into multiple if not atomic (don't allow partial to be possible)
  5. Okay to start new record midway through event list; okay to stop early?
    • drop subjects, attrition
    • “Year 2 replacement”; eg. record starts with MR2

Q&A w/ Doug F.

  1. Calculated fields/auto variables: useful, responsible for calculating chronotype for sleep study
  2. Setting RecordID: unsure, pugID (synonymous to lunaID) was inputted into RedCap by San Diego team. Specific # of IDs were generated (set #, no adding)
  3. Change question on an old instrument: has done this before, if the decision was made to no longer use a question, it was deleted. That data is found in past releses, but not used in future ones. (unsure if it deletes the old Q responses retroactively, asked but not entirely answered)
  4. Saving partially completed + redoing: Create space for repeat forms when it's anticipated… if you expect it to happen make one or two extra. [never has experienced this before]
  5. start/stop middawy through event? attrition? replacements?: label it as “incomplete” for attrition. doesnt use replacements so no insight. original cohort studied is only data sets used, if they drop out their events are simply incomplete.

FIONA server? they don't store any PIH (email/phone/anything..) it's in a separate system held by SD not pitt(still accessible by ABCD but not their own?) UPLOAD SURVEYS FROM CSV EXCEL FILE (horribly specific formatting) brnaching logic is probably the must usefuul thing (similar to SQIS(?) syntex) drawbacks/vulnerabilities: reports are clunky.. DF *thinks* qualtrics interfaces w/ redcap.. cannot say for certain but nonetheless

VL recap/takeaway: they basically supplied SD with the information they wanted, and then SD inputs (and continues to) the instruments/structural sides of it all. their RA's might know more about the RecordID/repeat visits.. might reach out to Andrew. i have more thoughts i just dont know them yet.