Many patients at risk for suicide are discharged from the ED with little or no intervention. Evidence-based suicide prevention interventions, like universal screening and safety planning should be adopted in clinical practice to help prevent suicidal behavior. This study will test the long-term sustainability of the nurse administered universal screening implemented in the original ED-SAFE study. Also, the investigators will test the impact of a new personalized Safety Planning Intervention guided by Lean has on suicide composite outcomes.

The ED-SAFE-2 will use a stepped wedge design where the original eight ED-SAFE sites will collect quantitative and qualitative data during the three phases: Baseline, Implementation, and Maintenance. Using this data, the ED-SAFE-2 will examine both within and between site differences for existing screening practices and new care processes, including safety planning. Most of the data collection on outcomes will be done by retrospective chart review.

A Lean Implementation Strategy will be used to ensure that adoption of improved care processes are fully supported vertically and horizontally within the organization, infrastructure is built that supports the efforts, and that the protocols fit naturally within roles, responsibilities, and clinical flow.

Consistent with the RFA's emphasis, the intervention target will be the clinician's behavior, including, at minimum, screening and safety planning. All emergency mental health and nursing personnel at the sites will be trained on safety planning, and Lean will be used to help ensure the safety planning is being implemented properly and consistently. The mechanisms of action of the combination of the safety planning training and Lean will be studied, allowing the team to establish both the effect the intervention has on the intervention target but also on the mechanisms of action comprised of departmental culture change and infrastructure support.

Suicide risk is much more prevalent among general emergency department (ED) patients than in the general community but this risk often goes undetected, especially among patients presenting with non-psychiatric complaints. By definition, universal screening is the only way to identify occult risk among patients presenting for non-psychiatric complaints. Until recently, however, little was known about how to implement universal screening in a clinical ED setting or whether doing so had any effect on detecting risk. The ED-SAFE has shown that it is feasible to implement universal screening and that the screening increases detection of suicide risk.

Simply identifying risk is not sufficient; efforts must be taken to mitigate risk and prevent suicidal behavior. However, many studies have shown that even those patients identified as having clinically significant risk are often discharged home without receiving any kind of active intervention during the ED visit, with many not even receiving a psychiatric evaluation.1-3 Brief interventions that are a good fit for the ED are needed. One such intervention has received strong research support and has already been accepted as a best practice in the United States Department of Veterans Affairs: Safety Planning Intervention.4 This intervention, however, has not been adopted in civilian EDs and little is known about how to effectively implement it, and whether doing so impacts suicide-related outcomes.

This study will address the following specific aims:

Aim 1: Test the long-term sustainability of nurse administered universal screening implemented in the original study across two new time periods. (a) The first is the period between the original ED-SAFE and the new study (ED-SAFE-2), which represents an ecologically valid "natural" state without any grant support, hereafter referred to as Baseline. (b) The second is the Maintenance phase of the new study, which will represent a time period spanning a minimum of four years after screening was initially implemented.

• Primary hypothesis: Sites that sustained high screening rates (intervention target) will sustain improved suicide risk detection (patient outcome) during each time period examined. Sustained screening practices will be mediated through ED organizational characteristics and enabling infrastructure (mechanisms of action).

Aim 2: Test the impact of implementing the new personalized Safety Planning Intervention for patients with suicide risk who are discharged from the ED.

• Primary hypotheses: Clinician training in safety planning, combined with implementation guided by Lean, is expected to increase safety planning by clinicians (intervention target), which will result in reduced suicide and suicide-related acute healthcare in the 6-months post-visit (suicide composite outcome). This will be more likely in sites with organizational characteristics and infrastructure that supports safety planning (mechanisms of action).

Aim 3: Test sustainability of safety planning during the Maintenance phase.

• Primary hypotheses: Sustained safety planning will result in sustained reductions in the suicide composite outcome. Sustained safety planning will be mediated by strong organizational characteristics and a robust enabling infrastructure supporting safety planning.

Inclusion Criteria:

• ED patients >18 years old
• Seen in ED triage

Exclusion Criteria:

• patients <18 years old
• not seen in ED triage
• Adults unable to consent
• Prisoners