At present, suicidal patients are still hospitalized to reduce their risk of committing a life-threatening act. Yet suicide research, long dominated by the ambition to identify people at risk, shows that the prediction model is largely ineffective. The time has come for a paradigm shift to: give up categorizing patients according to their risk; recognize that hospitalization, if not accompanied by a well-defined purpose, can become iatrogenic; formulate achievable short-term objectives, guiding such hospitalizations on the basis of a reproducible empirical model and demonstrating its benefits beyond keeping the patient alive. Interventions aim thus to reduce distress, personalize care and promote the change required by the suicidal crisis.