0655 Whose Mind after Midnight? Demographic and Clinical Variation in Nocturnal Suicide Risk

Abstract

Abstract Introduction Although suicides are more prevalent during the day, the risk of suicide is highest at night when adjusted for population wakefulness. It is unclear how nighttime patterns of suicide risk vary by demographic and clinical subgroups. Methods Population wakefulness data were combined with time of fatal injury data for 78,647 suicides from 2003-2017. The primary outcome was number of suicides per hour with wakefulness as an offset/exposure term. Subgroups included sex, race, ethnicity, military service, blood alcohol category and marijuana positive on autopsy, current depressed mood or partner conflict, and prior suicidal ideation or suicide attempts. Formal interaction tests identified significant subgroup variations adjusted for age, sex, race, and ethnicity where appropriate. Groups with significant interactions were subsequently examined using robust Poisson regression models to measure incident risk ratios (IRRs). Results There were subgroup interactions for age (p < 0.001), ethnicity (p = 0.014), blood alcohol level on autopsy (p < 0.001), and current partner conflict at time of fatal injury (p < 0.001). There was a sharp peak in risk around 4AM for individuals aged 15 to 24 (aIRR: 3.11 [2.39, 4.05]) and 25-34 (aIRR 2.15 [1.63, 2.86]), and at 6AM for individuals aged 65 to 74 (aIRR: 1.78 [1.40, 2.25]) and 75+ (aIRR: 2.25 [1.69, 3.00]). Hispanic individuals had an elevated risk between 10PM and 3AM with a maximal 232% increase at 2am (aIRR: 2.32 [1.67, 3.23]). Current partner conflict increased risk from 8PM to11PM and from 1AM to 3AM with a peak risk at 11PM (aIRR: 1.53 [1.24, 1.89]). Individuals with a BAL < 0.08 mg/dL showed a sustained increased risk for suicide between 8PM and 4AM; individuals with a BAL of 0.08 mg/dL or more showed greater risk (aIRR 2.33 [1.88, 2.89] at 1AM) than those with a BAL < 0.08 mg/dL (aIRR: 1.68 [1.40, 2.01] at 12AM). Conclusion Nighttime suicide risk varied significantly by age, ethnicity, blood alcohol level on autopsy, and current partner conflict. Military service was not associated with greater nighttime risk, which contrasts with prior findings. Support (if any)