Thursday, March 27, 2014
Present: Kairn Kelley, Rodger Kessler, Mark Kelly, Connie van Eeghen
Start Up: Welcome to Diantha Howard for a discussion about REDCap. Kairn finally received IRB approval – hooray!
1. Discussion: Advanced Topics on REDCap to follow previous CROW workshop
a. Randomization Module:
i. Allows user to input a coding order (the randomization plan), which REDCap uses to assign surveys. Randomization is a choice in Project Set Up (after allowing user permission to access Randomization under User Rights). Once the randomization plan is updated, it is fixed – it cannot be changed.
ii. Example: Kairn is planning to give 3 tests to subjects in 6 different possible orders. She would like to use all six orders, evenly distributed and randomly assigned. She has to develop the assignment system outside of REDCap, e.g. in Excel.
iii. Randomization may be stratified, for example, on gender, so that within each gender there will be a similar number of different treatments. In this case, there is a table for each stratified alternative, each with its own randomized plan.
iv. There are strategies for balancing randomization when the n is low, but these happen in the randomization plan outside of REDCap (e.g. an algorithm for a Latin Square).
b. Report building
i. Primary purpose: view the data – does NOT do much to summarize the data – in order to check for data collection or to produce a list of records for action
i. Keep in mind that there are two REDCap applications available to us: FAHC (behind a firewall) and UVM (the only application that can mail out surveys to email addresses)
ii. Multiple projects can be managed using the Bookmark feature in the future, after the next upgrade.
iii. The choice of a “longitudinal” study means using the EXACT same form (survey) more than once. It results in a file that creates multiple rows per subject.
d. Thank you, Diantha!
2. Next Workshop Meeting(s): Thursdays, 11:30 a.m. – 12:45 p.m., at Given Courtyard South Level 4. Remember: the first 15 minutes are for checking in with each other.
a. April 3: Rodger’s discussion of research design for PCORI study comparing integrated models of care.
b. April 10:
c. April 17:
Posted by Connie at 3/27/2014 04:19:00 PM