A significant number of people who attempt suicide and survive eventually die by their own hands, many within a year of the index attempt. A history of multiple past attempts further increases risk of eventual suicide. That most attempters do not later die by suicide is a statistical fact that should not distract psychiatrists and other mental health professionals from the substantial increase in risk associated with a suicide attempt. Short-term intensive treatment, often with psychiatric hospitalization, reduces immediate risk, but the standard of care often requires more than just a few days of generic inpatient care. Before discharging patients, the psychiatrist should be reasonably certain that the conditions associated with the attempt and initial suicide risk have improved in some significant and lasting way. Although for many patients, severe suicide risk is a relatively transient condition, patients should not be discharged just because they say they feel better or show superficial signs of lessened risk. Before sending the patient into the community, the psychiatrist should have good reason to believe that the dangerous condition(s) that precipitated the attempt and hospital admission have been ameliorated, and that the important improvements in the patient and his or her environment, on which the patient's safety relies, are both real and stable.