Skip to main content
Fig. 1 | BMC Medical Informatics and Decision Making

Fig. 1

From: Implementation and comparison of two text mining methods with a standard pharmacovigilance method for signal detection of medication errors

Fig. 1

Annotated dataset. Two MedDRA® HLGTs Device issues and Medication errors, and other product use errors and issues were applied as filters to the starting dataset of 154,209 narratives. This decreased the number of ICSRs to 25,328, and from the latter dataset a 10% random sample corresponding to 2533 ICSRs was taken for manual annotation. MedDRA® Medical Dictionary for Regulatory Activities, HLGTs high level group terms, ICSRs individual case safety reports

Back to article page