BACKGROUND : Suicide prevention is a core task in mental health services. Our objective was to determine whether Collaborative Assessment and Management of Suicidality (CAMS) reduced suicidal thoughts and behaviors and mental health distress more effectively than treatment as usual (TAU) in a heterogeneous patient population within specialized mental health care services. METHODS : In this observer-blinded pragmatic randomized controlled trial participants who scored 13 or above on Beck's Scale for Suicide Ideation-Current (BSSI-C) were included from seven in- and outpatient units. Primary outcome was suicidal ideation (BSSI-C). Secondary outcomes were mental health distress measured by the Outcome Questionnaire-45, and suicidal behaviors measured by the Suicide Attempt Self-Injury Count. Patients were assessed at baseline and after 6 and 12 months. RESULTS : The final intent-to-treat analyses included 78 participants (mean age 35.9 years, SD = 14.5, 41 females). The majority were depressed (65%), had a secondary diagnosis (73%) and 32% suffered from borderline personality disorder or borderline traits. After 6 months, CAMS participants reported lower levels of suicidal ideation compared to TAU (β = -4.29, 95% CI = -8.32 to -0.27, p = .036). Larger changes in mental health distress were observed for CAMS participants after 6 months (β = -11.87, 95% CI = -22.99 to -0.76, p = .036) and 12 months (β = -13.70, 95% CI = -24.88 to -2.51, p = .017). CONCLUSIONS : The modest sample size rendered the trial unable to detect small between-group differences. CONCLUSIONS : CAMS reduced suicidal ideation and mental health distress more efficiently than TAU in a heterogeneous patient population within specialized care.