Uncovering Annotation Inconsistencies in Death Investigation Notes to Improve Suicide Cause Attribution

Annotation inconsistencies in unstructured death investigation notes can lead to misattributed suicide causes, hindering effective suicide prevention strategies. This study proposes an empirical NLP approach to detect these inconsistencies, identify problematic instances, and demonstrate the effectiveness of correcting them.