Icon Legend

This session is not in your schedule.

This session is in your schedule. Click again to remove it.

Presentation Icons

Additional registration fee required

Faculty have requested this content not be shared on social media

CME Credit Offered

Poster Icons

Foundation Awardee

No Social Media

Back

Q&A

Computers

Poster Session 3

(684) Natural Language Processing Analysis of TikTok’s Most Popular #Pitocin Videos

Wednesday, February 14, 2024

8:30 AM – 10:00 AM

Location: Prince George's Exhibit Hall CDE, Lower Atrium Level

Primary & Presenting Author(s)

Bryan L. Aaron, MS (he/him/his)

Medical Student
University of Michigan Medical School
Ann Arbor, MI, United States

Coauthor(s)

KN

Katherine Neff, BA, MPH

Medical Student
University of Michigan Medical School
Ann Arbor, MI, United States
Fei Cai, MD

Maternal Fetal Medicine Fellow
Oregon Health & Science University
Portland, OR, United States
Luke P. Burns, MD

MFM Fellow
University of Chicago Medicine
Chicago, Illinois, United States

Objective: Among TikTok's 150 million American users, an estimated 90 million reportedly consult the app for healthcare advice before seeing a physician. Patients may therefore rely on TikTok as a source of information on intravenous (IV) oxytocin use in labor, however, no studies exist describing the tone of popular TikTok videos available on this topic. We, therefore, used a natural-language processing tool to evaluate the tone and sentiment of popular TikTok videos tagged #Pitocin.

Study Design:
The application Apify was used to compile information about the most-liked videos with the #Pitocin tag on April 26th, 2023. Two reviewers assessed 169 videos for eligibility and excluded 73 videos due to irrelevance or unavailability, resulting in 96 videos for analysis. We used a natural language processing tool, Valence Aware Dictionary for sEntiment Reasoner (VADER) to assess the polarity (positive/negative) and intensity (strength) of emotion of the video description text. The VADER score range was -1 to 1, with -1 indicating a highly negative sentiment, and 1 indicating a highly positive sentiment. Regression modeling was performed with VADER score as the exposure variable and the number of likes on the video as the response variable.

Results: The 96 reviewed videos with the #Pitocin tag had a total of 19,430,700 plays and 1,882,839 likes and averaged 60.4 seconds in length. 8 videos were removed from VADER analysis due to absent captions. Of the 88 video descriptions analyzed, 30 (34.1%) had values greater than 0, 23 (26.1%) had values less than 0, and 35 (39.8%) had values equivalent to 0. The adjusted R-squared value was equivalent to -0.002. VADER score did not significantly predict the number of likes on videos (β1= 7463, p-value = 0.89).

Conclusion:

More than a quarter of the most popular TikTok videos tagged #Pitocin had negative sentiments regarding IV oxytocin infusion, based on video descriptions. There was no statistical association between the popularity of a video and its sentiment score.