Predicting compassion fatigue among psychological hotline counselors using machine learning techniques

Area of Science:

• Psychology
• Clinical Psychology
• Mental Health

Background:

• The COVID-19 pandemic increased demand for psychological support, exposing hotline counselors to trauma, risking compassion fatigue.
• Compassion fatigue is a significant concern for mental health professionals, impacting their well-being and service delivery.
• Identifying predictors of compassion fatigue is crucial for developing targeted interventions for high-risk counselors.

Purpose of the Study:

• To explore the key predictors of compassion fatigue among psychological hotline counselors.
• To evaluate the effectiveness of various machine learning techniques in predicting compassion fatigue.

Main Methods:

• A cross-sectional study involving 712 psychological hotline counselors who completed validated questionnaires.
• Data analysis included chi-square tests for variable selection and multiple machine learning algorithms (logistic regression, decision tree, random forest, k-NN, SVM, Naïve Bayes) for prediction.
• Key variables measured were compassion fatigue, trait empathy, social support, mindfulness, self-efficacy, humor, life meaning, and post-traumatic growth.

Main Results:

• Meaning in life, counselor self-efficacy, mindfulness, and empathy were identified as the most significant predictors of compassion fatigue.
• Most machine learning models demonstrated good predictive performance, with Naïve Bayes achieving the highest AUC (0.803) and Random Forest showing the lowest classification error (23.64%).

Conclusions:

• Meaning in life, self-efficacy, mindfulness, and empathy are critical protective factors against compassion fatigue in hotline counselors.
• Machine learning techniques show promise for accurately predicting compassion fatigue, enabling proactive support for mental health professionals.