Though REDCap isn't a truly relational data system, it introduced (limited) one-to-many functionality in version 7.0, with "Repeatable Instruments" and "Repeatable Events." These are great features, but they may be confusing on first acquaintance.
The difference between "Repeatable Instruments" and "Repeatable Events" doesn't seem to be well documented, so here's my quick explanation of how they differ.
"Repeatable Instruments" is the simpler of the two. You don't need to enable "Longitudinal Data Collection" (i.e. "Events") in your project to use this feature. You just create the instrument, then enable "Repeatable instruments," then check the box to makes your selected instrument repeatable. You can now create as many different instances of this instrument as you like for the main record (usually a subject).
Here's an example. If you're working on a project that involves families seen at a single point in time, you could set up a regular (non-repeating) instrument ("Mother") to collect data on each mother, and then a repeatable instrument ("Child") to collect data on each child of that mother. Just create and fill out a new instance of the Child instrument for each child. The data can now be exported and analyzed in one-to-many fashion.
If you want to create "Repeatable Events," on the other hand, you must enable "Longitudinal Data Collection" as well as "Repeatable instruments and events." (Also, be sure to assign your instruments to events appropriately.) Now when you modify "Repeatable instruments and events," you can set one of the following options for each of your events:
- You can either leave the event totally non-repeating. (This is the default.)
- You can set selected instruments in that event as repeating. The effect here is much the same as setting up a "repeatable instrument" as described above.
- You can set the entire event as a repeating event. With this option, when doing data entry, instead of creating a new instance for each instrument in the event, you simply create a new event and all its instruments are repeated.
Why would you want to use repeatable events?
Recall that "events" are intended for longitudinal data entry - that is, data entry over time. Each instance of an event represents a repetition of a defined event at a new point in time.
Example: Imagine that you're doing a study in which subjects have some (defined or undefined) number of clinical encounters over a three-year period. At each of these encounters, you administer the same five assessment instruments. A convenient way to set this up in REDCap: create the clinical encounter as a repeating event and assign the five instruments to that event. Now, every time the subject comes in for an encounter, you just repeat the event; all the instruments will be automatically repeated and you can enter new data to each.
With that background, we can now address the original question: What's the difference between repeatable instruments and repeatable events?
- Repeatable instruments are simpler. You don't need longitudinal data collection.
- Repeatable events generally make sense for one-to-many relationships defined by different points in time (one subject to many ER visits, or many developmental points).
- Repeatable instruments generally make sense for one-to-many relationships that aren't defined by different points in time. (For example: one parent to many children; one subject to many medications; one subject to many labs.) But feel free to use a repeatable instrument for a time progression in a very simple project. (See for example the way subjects and exams are set up in our sample HPV Cytology project.)
- With a repeatable event, you repeat several instruments all at once.
- The data entry for the two looks a little different.
- The data exports look a little different.