APA

APA 2024: Investigating the use of stigmatizing language in clinical notes

May 13, 2024

Using a list of terms classified as stigmatizing by the National Institute of Drug Abuse, researchers developed a natural language processing (NLP) algorithm to extract these terms from clinical notes. Their findings highlight the need for modeling best practices to decrease stigma and bias, and further a need for provider education on what constitutes stigmatizing language (SL) in clinical documentation practices.

• Authors processed 980,194 notes for 2,700 patients with an opioid use disorder (OUD) diagnosis between 2010 and 2023 at a major academic medical center in New York City.
• The NLP system identified 2,279 (84.4%) patients with one or more SL term(s) in their notes. Terms such as "abuse", "opioid abuse", "addict", and "substance dependence" were the most frequent ones found in clinical notes.
• Patients in the age group 18 to 32 years had a lower incidence of SL in their notes compared with other age groups. Patients with a higher incidence of SL tended to be older, female, Black and/or Hispanic.
• Female clinicians used SL terms more frequently (50.0% in SL cohort vs. 42.4% in no SL cohort) than did male providers. Additionally, notes authored by mental health providers had a lower incidence of SL (11.6% in SL cohort vs. 25.6% in no SL cohort) compared with other specialties. Notes authored by social workers had a higher incidence of SL (6.0% in SL cohort vs. 0.2% in no SL cohort) in their notes.

Source:

Adekkanattu P, et al. Words matter: use of stigmatizing language in clinical notes of patients with opioid use disorder. Poster P05-090. Presented at: American Psychiatric Association Annual Meeting; May 4-8, 2024; New York City, NY. 2024. https://s7.goeshow.com/apa/annual/2024/poster_search.cfm?session_key=01141562-90B1-1C06-DFD2-AB78AEAB55A3&session_date=Sunday,%20May%2005,%202024