Predicting Changes in Mental Health Professionals' Clinical Practice Behaviors for Recognizing and Responding to Suicide Risk

Purpose: Clinicians who have primary contact with people at risk for suicide play a critical role in suicide prevention. Clinicians’ ability to accurately access an individuals’ suicide risk, which can be developed by a structured clinical training in evidence-based risk assessment, plays an important role in preventing suicide. Previous studies confirm the effectiveness of training for mental health professionals in improving suicide assessment and management skills. However, there is limited research on how clinicians’ attitudes and self-efficacy influence changes in practice behaviors over the training period. The majority of published studies utilize longitudinal, group-comparison statistical analyses, which are effective for testing for changes in key constructs (i.e., knowledge, attitudes, practice behaviors) but less so for predicting changes in key constructs. Extending research by employing statistical models to identify factors that influence practice behaviors will benefit the field of suicide prevention.

The Recognizing and Responding to Suicide Risk (RRSR) training is designed to improve clinicians’ competency to assess and respond to suicide risk. It is two-day skill-based training which teaches clinicians specific strategies to recognize and respond to client suicide risk. This study investigates if the RRSR training is effective in improving clinicians’ practice behaviors, including the ability to effectively assess and respond to suicide risk. Additional outcomes of interest are identifying factors that influence changes in practice behaviors.

Methods: Multilevel modeling (MLM) was used to examine the effectiveness of the RRSR training and identify key predictors of behavior change. Secondary data from a longitudinal study of the RRSR training was used (n=302). Standardized instruments were used to assess attitudes toward suicide prevention and self-efficacy; three detailed client vignettes with high inter-rater reliabilities were used to assess ability to assess and respond to suicide risk of clients.

Results: The unconditional model with ICC of .421 supported conducting MLM analyses. Significant changes were found in clinicians’ practice behaviors over time across the four models tested. Clinicians’ attitude, self-efficacy, gender, age, and educational degree were found to be significant predictors of clinicians’ practice behaviors: (1) attitudes toward suicide prevention (b =-.004, p =.001) and self-efficacy (b =.002, p =.004) in model 1; (2) gender (b =.04, p =.007) and educational degree (b =.03, p =.01) in model 2; (3) self-efficacy (b =.002, p = .04) and age (b =-.001, p =.03) in model 3; and (4) gender (b =.06, p =.03) and educational degree (b =.07, p =.002).

Conclusion and Implication: Findings from the current study are consistent with previous research, and support the RRSR as an effective training for improving clinicians’ practice behaviors even controlling for attitude, self-efficacy, and individual characteristics. Further, age, gender, and educational degree significantly predicted clinicians’ practice behaviors; gender and educational degree remained significant across models, including the random slopes and intercepts model. Trainees’ diverse backgrounds require customized training approaches; customized trainings may be even more effective to improve clinicians’ practice behavior and relatively reduce training expenses.