Table 4 Top 5 substance use-specific stigmatizing language in external dataset down-sampled/balanced

From: Detecting stigmatizing language in clinical notes with large language models for addiction care

Terms

Frequency Down-sampled

“alcohol abuse”

505 (12.19%)

“alcoholic cirrhosis”

438 (10.57%)

“tobacco abuse”

342 (8.25%)

“etoh abuse”

256 (6.18%)

“alcoholic”

253 (6.11%)

  1. Top 5 substance use-specific stigmatizing language from post-selected and labeled External Validation Down-Sampled. Note frequency is how many clinical notes this term was identified in. Some of these terms can be from the same note but are considered a unique tally in the frequency count.