More flexibility in cohort exit

[alondhe]

As @pbr6cornell outlined here: https://forums.ohdsi.org/t/limitations-of-atlas-cohort-definitions-tool/15752/2, we often need to have the cohort exit based on something other than end of obs period, continuous use of a drug (based on era), or a fixed duration.

I'm wondering if we can make this section more flexible to allow for a full set of criteria events and attributes, much like we have in the censoring.

[chrisknoll]

We'd just need to detail the desired functionality. Your example where you'd want to exit the cohort based on criteria is what censoring events does. Exiting on end of observation period + censoring events = exit cohort at a censor event. The reason why censoring events was provided separately from the other 'exit strategies' was that you may want to do end ob observation + censoring, or fixed duration + censoring, or continuous drug exposure + censoring.

I think that one new type of exit would be an exit based on era-fiying any set of domain criteria (the continuous drug exposure exit strategy assumes only drug exposure domain , with drug-exposure specific parameters such as persistence window and gap days). So, this is something I think has a clear objective and is something that would fit under a new 'exit strategy'.

For your use case, we'd just need to understand 'is it a new exit strategy we're thinning of?' and 'how should this exit strategy work?'