Analyzing Trends in Suicidal Thoughts Among Patients With Psychosis in India: Exploratory Secondary Analysis of Smartphone Ecological Momentary Assessment Data.

Background: India has the world's largest number of suicides, but there is little research on the trends in suicidal thoughts, especially for individuals with psychosis. More research is necessary to develop preventive interventions. Smartphone-based ecological momentary assessments (EMAs) can assess dynamic symptoms, but most EMA studies are conducted in higher income settings and have shorter (≤1 month) follow-up periods. Objective: This study aimed to examine the duration of onset to offset of suicidal ideation (SI) in tertiary hospital outpatients with psychosis in India. Methods: This study is an exploratory, secondary analysis of smartphone EMA data nested within the ongoing "Smartphone Health Assessment for Relapse Prevention (SHARP)" project. Tertiary hospital outpatients (n=50) with early course schizophrenia at 2 socioculturally different sites in India were recruited and given the "mindLAMP" app for monitoring mood through daily EMA surveys. The mood survey matched the 9-item Patient Health Questionnaire; the ninth item was used to define an instance of SI (score ≥1). A total of 14 patients with ≥1 SI instances who met the site-specific EMA survey use cutoff were included. We examined the between- and within-person variability in SI and computed the timescale of "episodic" SI (sequences of consecutive daily observations of SI score ≥1). Positive and Negative Syndrome Scale (PANSS) was used to assess changes in psychosis symptoms and its relationship with the temporality of SI. Results: Over approximately 11 (SD 2.1) months of EMA reporting on average, 3253 mood surveys were filled by the 14 participants (median 213, IQR 147-256). A total of 521 instances of SI were reported. Monthly SI instances showed substantial within- and between-person variations. Timescale summary statistics revealed episodic SI patterns in 11 patients, with an average of 5.9 episodes (SD 4.4; range:1-14; n=65) with an episode lasting on average 2.5 days (SD 1.5; range:1-5.3; n=27). There was an average lag of approximately 59, 66, and 81 days between the time of the first drop in PANSS positive, negative, and general psychopathology scores, respectively, and the last reported SI instance. Results after imputation of missing data showed an average of 12.1 episodes and 228 days (average lag) between the first drop in PANSS scores and last reported SI. This indicated that SI was an enduring vulnerability subsequent to the beginning of clinical improvement in psychosis. Conclusions: Our study adds to the much-needed evidence base in India to measure the dynamics of suicidal thinking within an individual, for more targeted preventive interventions. Further steps in EMA research are highlighted such as the use of higher frequency "burst" surveys to assess the duration of an SI episode in hours or minutes, and inclusion of both active and passive SI markers to measure the timescale of suicidal thinking.