4 Standard data mining runs
To assemble data mining counts and statistics that can be compared over time, JADER Signal Management includes data mining runs. All runs are refreshed as part of preparing a monthly Signal Management update.
The JADER signal management implementation uses the following types of data mining
runs:
- Data mining runs to produce disproportionality scores and report counts for the JADER data set that are incorporated into the signal summary table.
- Data mining runs to determine new or significantly changed cases relative to prior time periods.
- Data mining runs to produce disproportionality scores for the HLT and SOC levels of the MedDRA hierarchy.
- A 3D data mining run that is used to identify Drug-Drug-Event interactions that may merit further investigation.
Note:
Run descriptions in this document use the JADER Signal Management Configuration Attributes as installed. If attribute values, such as Strata Variables, are modified then the runs for the next refresh will be impacted accordingly.- Data mining runs produce disproportionality scores
JADER signal management information is based on nine data mining runs monthly refreshed to produce the disproportionality scores for drug-event terms. - Disproportionality Analysis Time Periods
To track changes to disproportionality scores over time, the signal summary table maintains disproportionality scores for five quarterly time periods. - Data mining runs for new or changed cases
Seven cumulative subset data mining runs are performed to determine new and significantly changed reports relative to prior time periods. - Data mining runs for SOC and HLT scores
Oracle Empirica Signal performs the following data mining runs to compute disproportionality scores at the SOC and HLT levels of the MedDRA event hierarchy. - Data mining run for viewing interactions
Oracle Empirica Signal executes a three-dimensional MGPS data mining run that is used to identify Drug-Drug-Event interactions that may merit further investigation.