Just-in-time Interventions for Reducing Short-term Suicide Risk

As the 12th leading cause of death in the United States, suicide is an alarmingly prevalent public health problem. The time period of highest risk for suicide and related behavior (SRB) is immediately (e.g., the weeks) following psychiatric hospitalization. Effective and scalable strategies for preventing SRB among individuals recently discharged from inpatient treatment are urgently needed. Encouragingly, brief evidence-based interventions now exist that reduce risk for suicide after acute psychiatric care. A primary example is safety planning, which involves developing a prioritized list of coping strategies and sources of support that people can use to mitigate future suicidal crises; this intervention is now recommended as part of standard inpatient clinical care. However, it has been shown that well over one-third of suicidal patients who have a safety plan never use it. Additionally, little is known about when (and for whom) specific components of the safety plan, which includes internal coping strategies, social support activities, and help-seeking behaviors, are most accessible and effective. Optimizing real-world safety plan use with scalable intervention strategies delivered via mobile technology has great potential to improve the effectiveness of an already promising intervention and advance precision treatment for high-risk individuals.

This project will employ the recently developed micro-randomized trial (MRT) design to test the effectiveness of brief, just-in-time interventions aimed to promote real-world use of the safety plan and its specific components. Both the method (e.g., phone call or text messaging by a human or automated forms of smartphone-based messaging) and content (e.g., recommendation to use the safety plan in its entirety or a specific component) will be randomized according to individuals' current levels of suicidal urges and intent. Participants will be psychiatric inpatients admitted for suicidal thoughts or behaviors who agree to participate in an intensive longitudinal monitoring protocol involving real-time smartphone-based surveys for the 28 days after hospitalization; after each completed survey, participants will be repeatedly "micro-randomized" to one of several just-in-time interventions. The investigators will test the hypotheses that the brief just-in-time interventions will be both acceptable and feasible, and associated with increased use of safety plan components (target mechanism) and reductions in suicidal thoughts (proximal outcome) at varying levels of suicidal thoughts, and that intervention effects will vary by both intervention method and content. The investigators will also explore internal and external contextual moderators (e.g., affect, social support) of proximal intervention effects, and collect qualitative data that will inform intervention refinement as well as the future development and implementation of just-in-time adaptive interventions.

The investigators will first conduct a small pilot MRT (N=10 participants), after which qualitative feedback collected on the intervention methods, content, amount of support, timing, and triggering, as well as overall data on acceptability and feasibility, will be used to refine the interventions before the full MRT (per Question #1 outlined above). The full MRT of the refined just-in-time interventions will be conducted in N=175 participants (and answer Questions #2 and #3 above).

Inclusion criteria

• Experiencing suicidal thoughts as part of their inpatient admission
• Access to a smartphone following discharge
• Ability to speak and write English fluently

Exclusion criteria

• Any factor that impairs the ability to effectively participate in the study (e.g., significant cognitive impairment, intellectual disability, presence of violent behavior, psychotic illness/symptoms determined by the treating clinician to impair ability to understand the study or provide informed consent)
• Failure to correctly answer all true/false questions in the consent form