Multi-informant prediction of near-term suicidal behavior independent of suicidal ideation

Abstract

Background: Both the Suicide Crisis Syndrome (SCS) and clinicians' emotional responses to suicidal patients are predictive of near-term suicidal behaviors. Thus, we tested predictive validity of a combination of the proposed Diagnostic and Statistical Manual SCS criteria and the Therapist Response Questionnaire Suicide Form (TRQ-SF) for near-term suicidal behavior.

Methods: The presence of SCS in adult psychiatric outpatients (N=451) was assessed using relevant items from validated psychometric assessments. Clinicians completed the TRQ-SF immediately after patient intake. Suicide attempts (SA) and a combination of suicide plans and attempts (SPA) were measured at one month follow-up (N=359).

Results: At follow-up nine patients reported having SPA and seven reported SA. Meeting the SCS criteria were associated with near-term SA (χ2=5.987, p<0.01), while high TRQ-SF scores were associated with both near-term SA (χ2=5.971, p<0.05) and SPA (χ2=7.069, p<0.01). Meeting either the SCS or having high TRQ-SF scores, but not both, was associated with near-term SA (χ2=11.893, p<0.01) and SPA (χ2=11.449, p<0.01). Incremental predictive validity over standard suicide risk factors and individual scales was demonstrated in logistic regressions.

Conclusions: Multi-informant risk assessment not reliant on patient self-reported ideation appear to enhance predictive power of traditional risk assessments in identifying imminent suicide risk.