Emergency Department Safety Assessment and Follow-up Evaluation (ED-SAFE)

ED-SAFE was a multicenter study of suicide prevention. Please contact Drs. Ed BoudreauxCarlos Camargo or Van Miller if you would like additional details.

Example publications:
Boudreaux ED, Camargo CA Jr, Arias SA, Sullivan AF, Allen MH, Goldstein AB, Manton AP, Espinola JA, Miller IW. Improving suicide risk screening and detection in the emergency department. Am J Prev Med 2016; 50: 445-453. PMID: 26654691

Miller IW, Camargo CA Jr, Arias SA, Sullivan AF, Allen MH, Goldstein AB, Manton AP, Espinola JA, Jones R, Hasegawa K, Boudreaux ED. Suicide prevention in an emergency department population: the ED-SAFE study. JAMA Psychiatry 2017; in press. PMID: 28456130

Co-Principal Investigators
Edwin Boudreaux, PhD, University of Massachusetts Medical School, Worcester, MA 
Carlos A. Camargo, MD, DrPH, Massachusetts General Hospital, Boston, MA
Ivan Miller, PhD, Butler Hospital, Providence, RI

Funding Source: National Institute of Mental Health (1U01MH088278)

Purpose: (1) To develop and test a standardized approach to screening emergency department (ED) patients for suicide risk using a tool suitable for systematic use in general medical EDs; (2) To refine and test an ED-initiated intervention to reduce suicidal behavior and associated morbidity and mortality among people who self-identify or screen positive for suicidal ideation; and, (3) To complete a comparative economic analysis of treatment as usual, screening, and the intervention.

Rationale: The ED is an opportune setting for suicide prevention, with >500,000 ED visits related to intentional self-harm in 2006. These efforts should begin with a study of whether universal screening, brief intervention, and referral to treatment is effective at improving suicide detection, enhancing quality of care, and reducing suicidal behavior. There is ample opportunity to identify and intervene with suicidal patients, with at least 3% having had thoughts of killing themselves within the two weeks before their visit. Despite the public health significance of suicidal behavior, there have been relatively few controlled trials which evaluated interventions to reduce suicidal behavior.

Procedure: The study consisted of 3 phases: (1) Treatment as Usual; (2) Screening Alone; and (3) Intervention. Each of the 8 sites staffed the ED with research assistants (RAs) who prospectively enrolled a sample of 60 participants into each phase. RAs staffed the ED for 10-hour shifts during the peak volume hours of 12 noon to 10 PM for five days per week. Approximately half (n=30) were patients who had attempted suicide and half (n=30) were patients who had active suicidal ideation (but no recent attempt). This yielded a total of 480 prospectively enrolled participants for each phase, or a combined total of 1,440 participants across all three phases. Participants enrolled regardless of whether they are admitted to an inpatient facility or discharged from the ED. Participants enrolled until each strata (attempters vs. ideators) was filled.

Inclusion criteria:

  • Age ≥18 years
  • Thoughts of killing oneself in the past 2 weeks
  • An actual, aborted, or interrupted attempt to kill oneself in the past 2 weeks, including current visit
  • Able to consent (alert, fully oriented, not intoxicated, able to paraphrase the study requirements)
  • Willing to complete telephone follow-up assessments

Exclusion criteria:

  • Otherwise medically or cognitively unable to participate in the assessment or counseling (e.g., sustained altered level of consciousness, psychosis, hostile behavior, victim of sexual assault, intubation, persistent vomiting, severe pain)
  • Out-of-state residents
  • Currently in state custody
  • Pending legal action
  • Lack of permanent residence
  • Lack of reliable telephone service
  • Non-English or non-Spanish speaker
  • Already enrolled into the ED-SAFE trial

Subjects completed several assessments while in the ED and then received follow-up telephone calls 6, 12, 24, 36, and 52 weeks after the ED visit from a centralized call center at University of Massachusetts. Healthcare utilization, and suicidal ideation and behavior were assessed during these calls. In addition, all participants were followed for 12 months using death registries and electronic administrative databases associated with the healthcare system with which the site is affiliated.

During the Treatment as Usual phase, potential subjects were identified during the course of usual care. During the Screening Alone phase, a standardized screener completed by the primary nurse as part of routine care, not simply for research case identification purposes, and occured regardless of whether research staff is present. During the Intervention phase, each ED implemented a multi-component, systems-based Intervention called the Safety Care-chain. The Safety Care-chain combined elements of: (a) Question, Persuade, Refer (QPR) intervention administered by nursing staff in the ED, and (b) a modified version of the Coping Long Term with Attempted Suicide Program (CLASP-ED), a series of up to 7 semi-structured telephone counseling calls to the patient and 4 to “significant others” over the 12 months after the ED visit done at a centralized call center at Butler Hospital. QPR will be implemented universally to all suicidal patients, regardless of whether they are ultimately enrolled into the trial, as part of a comprehensive suicide management protocol (i.e., it is a systems-based change). The CLASP-ED telephone counseling calls were administered only to study subjects.

A random chart review of 480 patients for each phase was used as a method of assessing whether universal screening actually occurred.

Outcomes: We examined whether universal screening increases detection of suicide, and whether screening and the intervention decrease suicidal ideation and behavior in the 12 months following the ED visit. In addition, a comparative economic analysis was completed which quantified costs associated with each phase and computed quality adjusted life years (QALYs) and preference-based health-related quality of life (HRQOL). Finally, healthcare provider self-reported suicide knowledge, screening, and management practices were assessed via survey before and after the implementation of universal screening (and after the data collection is complete for the entire study).