Onset and Persistence of NSSI Patterns and risk of Suicide Reattempt: Results from the SURVIVE Study
M. Arqueros, E. Suarez-Soto, M. Diaz-Marsa, A. De La Torre-Luque, A. Cebria, M. Elices, M.-A. Boti, P. Lopez-Peña, A. Palao-Terrero, M. Ruiz-Veguilla, A. Garcia-Fernandez, T. SURVIVE Consortium, V. Perez-SolaIntroduction
After a suicide attempt, reattempt risk is concentrated in the first year, yet how post-attempt non-suicidal self-injury (NSSI) patterns relate to the timing of reattempts remains unclear. Hazard-based models can quantify this association.
Objectives
To evaluate whether NSSI pattern membership predicts time to suicide reattempt over 12 months in a multivariable Cox model.
Methods
Adults with a recent suicide attempt (n=685) were assessed ≤15 days post-attempt and followed for 12 months. NSSI pattern (reference=NN: No→No) was coded as Onset (No→Yes), Remission (Yes→No), or Persistence (Yes→Yes). The initial covariates were age, sex, number of prior attempts, BIS-11 subscales, ACSS-FAD, MINI diagnoses (count), BSI-GSI, reasons for ideation, and NSSI pattern. Backward elimination by Akaike’s Information Criterion (AIC) yielded the final model. We report hazard ratios (HR) with 95% CIs.
Results
The final Cox model (Figure 1) retained age, number of prior attempts, reasons for ideation (reference=External), and NSSI pattern (
AIC
=833.7;
χ²
(7)=22.26,
p
<.002; concordance=.632). Older age was associated with lower hazard (
HR
=0.98, 95% CI [0.96, 0.99]); more prior attempts with higher hazard (
HR
=1.02, 95% CI [1.01, 1.04]). Relative to NN, Remission (
HR=
0.56, 95% CI [0.32, 0.96]) and Persistence (
HR
=0.58, 95% CI [0.35, 0.96]) were associated with reduced 12-month hazard; Onset did not differ significantly.
Image 1: Long description.