This topic will track discussion of updating the pharmacovigilance signals imported into LAERTES
-
Milestone target date: 1/31/2014
-
Milestone description:
- generate association stats for all genereic-MedDRA PT pairs we can and load them into the evidence base served by LAERTES
-
Requirements
-
Format of signal data provided to Rich for loading into the evidence base (the format I suggested only makes sense for positive cases. The provenance data for the dataset will be useful for both positive and negative cases.)
omop_drug_concept_id|rxnorm_concept_name|omop_outcome_concept_id|meddra_pt_concept_name|aers_ebgm|aers_eb05|aers_eb95|aers_PRR|aers_ROR|aers_date_signal_emerged|aers_date_first_reported|aers_current_report_count -
As a best practice, LAERTES will track the provenance of the included evidence sources. Would you or someone on your team also please provide the following metadata on the source of the evidence:
a. the tool/method used to create the data
b. the original source of the data (Just a link to the FAERS files downloaded or the download site locate would be sufficient)
c. the time at which the data was obtained/generated/revised
d. the authors or publishers of the data
e. the license associated with the data
f. prior versions of the data (None at this time)
-
-
Questions to be addressed
-
Should the dataset include all generic drug-meddra pt pairs or only those pairs with at least k AERS
reports (e.g. k=10/20/50)? Alternatively, we could do something like: find all X drugs which must have >10000 exposures in at least one database (CCAE, MDCR, MDCD, GE) . then another filter on all Y conditions with >10000 occurrences in one database. then, the matrix of interest is X*Y…what do you think?