(PS13-24) Rest-activity Rhythm Irregularity Is Associated with Suicide Risk in a High-risk Adolescent Community Treatment Sample

Abstract Text:

Background:  Sleep and circadian disturbances have been identified as a promising potential  modifiable indicator of near-term suicide risk, with studies demonstrating prospective associations between disturbances and next-day, next-week, and next-month suicidal thoughts and behaviors (STBs) in youth populations. However, no research to date has examined the near-term associations between rest-activity rhythms (RAR), an actigraphy-based measure of circadian rhythm regularity, and STBs in youth. The current project aimed to address this gap by examining the associations between RARs and same-week and next-week STBs in adolescents attending a community intensive outpatient program for STBs.

Method:  Participants are 58 adolescents (mean age 17.06; 72.4% female sex at birth) who participated in the Sleep Predicting Outcomes in Teens (SPOT) study. Participants wore an actiwatch (GT9X Link actigraph) for up to 3 months, from which RAR regularity metrics--interdaily stability (IS) and intradaily variability (IV)--and daytime activity levels (M10) were computed on weekly intervals. Weekly clinician ratings of depression and suicidal ideation severity were derived from the Adolescent Longitudinal Follow-Up Evaluation (ALIFE) Psychiatric Status Rating (PSR) scales. Suicidal behaviors were not examined in analyses due to low rate of occurrence. We employed mixed level models to describe the patterns of RARs over time, and to examine concurrent and prospective associations between weekly RARs and suicidal ideation (both the odds of having suicidal ideation with a specific method via logistic regression, and ideation severity via linear regression). 

Results:  Intra-class coefficient analyses indicated significant within and between-person variability in RARs and daytime activity levels week-to-week (ICCs=0.51—0.66). Lower IS was associated with higher odds of occurrence of suicidal ideation with method (OR=0.61 [95% CI=0.41—0.92, p=.018) and more severe suicidal ideation ratings (β=-0.15, SE=0.05, p=.006) during the same week. Lower IS was also associated with higher odds of next-week suicidal ideation occurrence (OR=0.52 [95% CI=0.29—0.91], p=.023). Moreover, within-person week-to-week decreases in IS (β=-0.13, SE=0.06, p=.019) and M10 (β=-0.11, SE=0.05, p=.031) were associated with more severe suicidal ideation ratings during the latter week. These associations remained significant even after covarying for depression severity and previous-week SI ratings. 

Conclusion:  Our results indicate greater 24-hour irregularity in RARs may be predictive of near-term suicide risk in adolescents. Considering RARs are modifiable by behavioral interventions, future research may benefit from considering 24-hour circadian metrics, in addition to solely sleep-focused variables, in the study of sleep and suicide risk identification, prevention, and treatment in adolescents.