Abstract: TH-PO010
Identification of Topics from Trainee Posts on Nephrology from Public Forums
Session Information
- AI, Digital Health, Data Science - I
November 02, 2023 | Location: Exhibit Hall, Pennsylvania Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Augmented Intelligence, Digital Health, and Data Science
- 300 Augmented Intelligence, Digital Health, and Data Science
Authors
- Dai, Yang, Icahn School of Medicine at Mount Sinai, New York, New York, United States
- Borvick, Miriam S., University of Nevada Reno, Reno, Nevada, United States
- Ehrenfeld, Ricki, Touro University, New York, New York, United States
- Nadkarni, Girish N., Icahn School of Medicine at Mount Sinai, New York, New York, United States
- Chan, Lili, Icahn School of Medicine at Mount Sinai, New York, New York, United States
Background
Interest in a career in nephrology has waned. Career perceptions that are freely shared on social media sites may influence a trainee’s decision when choosing a subspecialty. Examining topics of discussion on public forums have not previously been done.
Methods
We extracted threads from a popular online forum, the student doctor network (SDN), with titles that contained the keywords Nephrology, Nephrologist, Nephro, Dialysis, Kidney, and Renal. We removed posts with <20 words. We performed topic modeling using BERTopic. BERTopic is a topic modeling technique in the Python library that combines transformer embeddings and clustering model algorithms to identify topics using natural language process. We then manually reviewed three posts per topic to better characterize the topics. Additionally, two authors conducted a manual review of 1000 posts to gauge the underlying sentiments expressed. When disagreement occurred, a third reviewer was asked to rate the post and the most frequent category was used.
Results
We included a total of 1725 posts in our final analysis. BERTopic identified 23 topics (Figure 1A). The topics with the largest number of posts were “Nephrology responsibilities and reimbursements”, “Fellowship programs mislead”, and “Academic vs private practice nephrology”. The top 10 topics are visualized in 2D space in figure 2B, where each point represents a post, and topics are separated by color. With manual review, 46% were classified as negative while only 6% were classified as positive. The remaining posts were unclear or irrelevant.
Conclusion
Discussions regarding nephrology careers are present on social media. Topics vary from workload to financial compensation. Unfortunately, a large proportion of posts are negative. To improve the nephrology workforce, we must address the concerns of trainees identified here.
Funding
- NIDDK Support