JMIR Publications

ABSTRACT

Background:

Suicide is the 12th leading cause of death in the U.S. Healthcare provider training is a top research priority identified by the National Action Alliance for Suicide Prevention; however, evidence-based approaches that target skill-building are resource intensive and difficult to implement. Novel computer technologies harnessing artificial intelligence are now available that hold promise for increasing the feasibility of providing trainees opportunities across a range of continuing education contexts to engage in skills practice with constructive feedback on performance.

Objective:

The aims of this pilot study were to evaluate the feasibility and acceptability of an eLearning training in suicide safety planning among nurses serving patients admitted to a U.S. level 1 trauma center for acute or intensive care. The training included a didactic with demonstration, practice of micro-counseling skills with a web-based virtual patient (“Client Bot Emily”), role-play with a patient actor, and automated coding and feedback on general counseling skills based on the role-play via a web-based platform (“Lyssn Advisor”). Secondarily, we examined learning outcomes of knowledge, confidence, and skills in suicide safety planning descriptively.

Methods:

Acute and intensive care nurses were recruited between November 1, 2021 and May 31, 2022 to participate in a formative evaluation using pre-training, post-training, and 6-month follow-up surveys as well as observation of nurse performance delivering suicide safety planning via standardized patient role-plays over 6 months and rated using the Safety Plan Intervention Rating Scale (SPIRS). Nurses completed the System Usability Survey (SUS) after interacting with Client Bot Emily and reviewing general counseling scores based on their role-play via Lyssn Advisor.

Results:

Eighteen nurses participated; the majority identified as female (n = 17, 94.4%) and White (n = 13, 72.2%). Of the 17 nurses who started the training, 82% (n = 14) completed it. On average, the SUS score for Client Bot Emily was 70.3 (SD = 19.7) and for Lyssn Advisor was 65.4 (SD = 16.3). On average, nurses endorsed a good bit of knowledge (M = 3.1, SD = 0.5) and confidence (M = 2.9, SD = 0.5) after the training. After completing the training, no nurses scored above the expert-derived cut-off for proficiency on the SPIRS (≥ 14); however, on average, nurses were above the cut-offs for general counseling skills per Lyssn Advisor (for Empathy M = 4.1, SD = 0.6; for Collaboration M = 3.6, SD = 0.7).

Conclusions:

Findings suggest the completion of the training activities and use of novel technologies within this context are feasible. Technologic modifications may enhance the training acceptability and utility, such as increasing the virtual patient conversational abilities and adding automated coding capability for specific suicide safety planning skills. Clinical Trial: ClinicalTrials.gov registration identifier: NCT05178121