search

ohdsi-studies / PioneerWatchfulWaiting

This study is part of the joint PIONEER - EHDEN - OHDSI studyathon in March 2021, and aims to advance understanding of clinical management and outcomes of watchful waiting in prostate cancer.
Apache License 2.0
7 stars 18 forks source link

Fix negative cohort periods #87

Closed MaximMoinat closed 3 years ago

MaximMoinat commented 3 years ago

Fixes #84

Better fix than proposed in #86, as this applies the offset index date at the beginning and the rest of the cohort sql definition uses this new date. Inclusion criteria and censoring events should be all relative to the offset index date.

keesvanbochove commented 3 years ago

This looks better from a programming perspective, but the problem is that if the inclusion criteria are relative to the new offset index date, most cohorts will end up being empty... because for instance one criterium is 'no history of PCa', and if you look back from the offset date you will by definition find one, namely the original cohort entry event of having a first ever PCa diagnosis. Unless I'm missing something, I am testing it out on some synthetic data.

MaximMoinat commented 3 years ago

Agreed, this whole offset thing is much more complex than we anticipated.