In all of these notebooks, I'm currently conflating the cohort begin/end date with the observation window. That's not correct. Ideally we could tunnel the observation begin/end as global variables, such that low-level filters could transparently use observation begin/end if it happens to be set. This is independent of export, since we need the notebook calculations to respect this as well, in this way export won't be materially different.
Cohort begin/end date should be used for hospitalizations or other reentrant cases where a subject enters/exits the cohort multiple times.
In all of these notebooks, I'm currently conflating the cohort begin/end date with the observation window. That's not correct. Ideally we could tunnel the observation begin/end as global variables, such that low-level filters could transparently use observation begin/end if it happens to be set. This is independent of export, since we need the notebook calculations to respect this as well, in this way export won't be materially different.
Cohort begin/end date should be used for hospitalizations or other reentrant cases where a subject enters/exits the cohort multiple times.