search

oxford-pharmacoepi / MegaStudy

4 stars 2 forks source link

Possible duplicates in incidenec and DUS results #74

Open lilliniroberto opened 3 months ago

lilliniroberto commented 3 months ago

Dear colleagues, after checking the results for incidence/prevalence and DUS, we have found some possiblke duplicates:

While we go on checking to be sure that this really happened, how can we control for this by the analysis packages? This because we are wondering if this could be an issue for the results.

Thank you in advance Roberto

tiozab commented 3 months ago

@lilliniroberto you are saying that you have two databases in one cdm and therefore some of the patients are duplicated? do they have the same person_id or they have different person_id but you are sure they are the same person?

lilliniroberto commented 3 months ago

We have two different sources imported in drug_exposure of the CDM, which might report the same information about drugs administration to the same patient in the same date. We are sure that the person_id is the same.

tiozab commented 2 months ago

@lilliniroberto sorry for my late reply, I was on holiday. The package does not control for this (it expects a clean dataset). So any data management has to be done before the analysis stage.