Suicide and Self-Injury
Assessing suicidality in the moment: acceptability of a smartphone-delivered suicide EMA to partial hospital patients.
Erin E. Beckham, B.A.
Clinical Research Assistant
McLean Hospital
Belmont, Massachusetts
Heather Martin, None
Research Assistant
McLean Hospital
Belmont, Massachusetts
Yas Salon, None
Research Assistant
Tufts University
Belmont, Massachusetts
Melanie Hom, Ph.D.
Clinical Assistant Professor
Stanford University
San Jose, California
Courtney Beard, Ph.D.
Associate Professor
McLean Hospital
Belmont, Massachusetts
The month post-discharge from acute psychiatric care is a high-risk time for re-hospitalization and suicide. Despite the importance of assessing for risk during this critical period, most measures of mood and suicidality rely on retrospective self-report ratings. To address the issue of recall biases, researchers are increasingly using Ecological Momentary Assessment (EMA) to capture moment-to-moment fluctuations in psychological constructs, including suicidal thoughts and behaviors. This study examined the acceptability of a smartphone-delivered EMA survey assessing shifts in suicidal ideation and behaviors among individuals traversing the challenging transition from acute psychiatric care.
In the 30 days following discharge from a partial hospital program, patients (N=20) were prompted with brief EMA surveys four times daily that assessed suicidal thoughts and behaviors (SI-EMA). At one-month post-discharge, patients completed a semi-structured qualitative interview that probed participant experiences of being asked about suicide multiple times daily, the saliency of study monitoring procedures (i.e., responses not monitored), and preference for smartphone or face-to-face suicide risk assessment. Two authors reviewed interview transcripts and developed a coding framework, which was reviewed with the senior author. Two authors then separately coded responses and met with the study team to resolve discrepancies.
Overall participants reported answering questions honestly (M=4.3 on a 1-5 scale). Reasons for inaccurate response included “fear of hospitalization/judgement”, “feeling distressed”, and “answering on autopilot”. Participants reported positive (i.e., “awareness and mindfulness”), neutral (i.e., “fine”), metacognitive (i.e., thinking about the assessment affecting responses), and negative (e.g., finding surveys “anxiety provoking”, “disturbing/heavy”, “forceful/intrusive”, “annoying”, “triggering/reminder”) emotional reactions. 12 participants reported preferring SI-EMA, as they found the format “less intense/awkward”. 3 participants preferred in-person assessments as they were “overwhelmed by the text” and “preferred having a clinician present as a resource”. Despite being told by study staff and in-app prompts that SI-EMA responses were not being monitored, 8 (44.4%) participants still reported feeling monitored. Participants expressed a range of views around the ethics of monitoring of SI-EMA responses, including how often and the risk level at which these responses should be monitored.
Our qualitative analysis indicates heterogeneity of experiences with SI-EMA. While the assessment was acceptable to many participants, some participants also reported significant distress from answering suicide prompts multiple times daily. Results show that while many people prefer the anonymity of SI-EMA, individual differences in emotional reactions and monitoring perceptions highlight the importance of tailoring assessment format to the patients’ preferences. Thoughtful consideration of risks and benefits is indicated when clinicians and researchers utilize SI-EMA.