APS
2026 APS Annual Convention · 2026
Identifying Latent Classes of Repeat Crisis Line Users: A Natural Language Processing Approach to Early Detection
- Tomas Madeira
- Greg Buda
Crisis Text Line - Elizabeth Olson
Crisis Text Line - Kelly Zuromski
Crisis Text Line
Abstract
Repeat texters are a large, high-risk group of crisis service users. To understand this population better, we analyzed 100,000 initial conversations using natural language processing and clustering methods. Preliminary results suggest distinct patterns in topics discussed in the first conversation, which could inform tailored interventions for this heterogeneous population.
