When transforming the REDCap data to our representation, so far we’ve been using participant_id + event_id as the PK. So the timeline was given by the events.
Now, it looks like we’re moving to an organisation where the timeline will be given by the study weeks. How do we want to represent this in the data? I assume have participant_id + week(_id) as the PK. [Question 1️⃣] But do we still want to include event for information? Can this extra categorisation ever be useful to analysts?
Now to actually add the week info, we need to be able to determine what week each row in the data corresponds to. In the majority of cases, we can already do this: any row coming from a visit will have the week of that visit and any row coming from a week-specific form will have that week. However, we have a handful of other forms where getting the week is not so straightforward. As far as I can see, these usually have a date field, so we could get the week based on that.
[Question 2️⃣] Should I start working on extracting the week info like this or should I hold off because we expect to get a nice mapping? (This overlaps with some of Signe's questions in #183 .)
@K-Beicher @lwjohnst86
When transforming the REDCap data to our representation, so far we’ve been using
participant_id+event_idas the PK. So the timeline was given by the events.Now, it looks like we’re moving to an organisation where the timeline will be given by the study weeks. How do we want to represent this in the data? I assume have
participant_id+week(_id)as the PK. [Question 1️⃣] But do we still want to includeeventfor information? Can this extra categorisation ever be useful to analysts?Now to actually add the week info, we need to be able to determine what week each row in the data corresponds to. In the majority of cases, we can already do this: any row coming from a visit will have the week of that visit and any row coming from a week-specific form will have that week. However, we have a handful of other forms where getting the week is not so straightforward. As far as I can see, these usually have a date field, so we could get the week based on that.
[Question 2️⃣] Should I start working on extracting the week info like this or should I hold off because we expect to get a nice mapping? (This overlaps with some of Signe's questions in #183 .)
@K-Beicher @lwjohnst86