BACKGROUND: Brief interventions for suicide risk among patients treated in acute care settings like the emergency department are needed. The Safety Planning Intervention is a promising approach but has yet to undergo a high quality, individual level randomized controlled trial. PURPOSE: This paper describes the methods associated with an individual level randomized controlled trial of the Safety Planning Intervention compared to a control condition comprised of reviewing risk factors and warning signs. METHODS: The sample comprised patients 18 years and older presenting to one of three different emergency departments with suicide related emergencies (target n = 484). Eligible patients were approached, consented, and randomized to the intervention (Safety Planning Intervention) or control (risk factors and warning signs). They were assessed at 1, 3 and 6 months after their index visit. The primary outcome is suicidal behavior. The study also assessed mechanisms of action. Data analyses are pending. CONCLUSIONS: We identified and addressed key challenges to studying suicidal patients in the emergency department, including difficulty enrolling during the emergency department visit, ascertaining outcomes in patients that are historically very difficult to follow, and addressing the ambiguity of suicidal behavior. ClinicalTrials.gov Identifier: NCT03227991.

BACKGROUND: Safety planning is a brief intervention that has become an accepted practice in many clinical settings to help prevent suicide. Even though it is quick compared to other approaches, it frequently requires 20 min or more to complete, which can impede adoption. A self-administered, Web-based safety planning application could potentially reduce clinician time, help promote standardization and quality, and provide enhanced ability to share the created plan. OBJECTIVE: The aim of this study was to design, build, and test the usability of a Web-based, self-administered safety planning application. METHODS: We employed a user-centered software design strategy led by a multidisciplinary team. The application was tested for usability with a target sample of suicidal patients. Detailed observations, structured usability ratings, and Think Aloud procedures were used. Suicidal ideation intensity and perceived ability to cope were assessed pre-post engagement with the Web application. RESULTS: A total of 30 participants were enrolled. Usability ratings were generally strong, and all patients successfully built a safety plan. However, the completeness of the safety plan varied. The mean number of steps completed was 5.5 (SD 0.9) out of 6, with 90% (27/30) of participants completing at least 5 steps and 67% (20/30) completing all 6 steps. Some safety planning steps were viewed as inapplicable to some individuals. Some confusion in instructions led to modifications to improve understandability of each step. Ratings of suicide intensity after completion of the application were significantly lower than preratings, pre: mean 5.11 (SD 2.9) versus post: mean 4.46 (SD 3.0), t27=2.49, P=.02. Ratings of ability to cope with suicidal thoughts after completion of the application were higher than preratings, with the difference approaching statistical significance, pre: mean 5.93 (SD 2.9), post: mean 6.64 (SD 2.4), t27=-2.03, P=.05. CONCLUSIONS: We have taken the first step toward identifying the components needed to maximize usability of a self-administered, Web-based safety planning application. Results support initial consideration of the application as an adjunct to clinical contact. This allows for the clinician or other personnel to provide clarification, when needed, to help the patient build the plan, and to help review and revise the draft.