Frequently Asked Questions

The main higher-risk patterns identified were alcohol use disorder with acute alcohol use during the index suicide attempt, more than 2 suicide attempts in the past 3 years, and anxiety disorder in a specific lower-attempt subgroup. In the survival tree analysis, 38 of 127 patients with alcohol use disorder and acute alcohol use during the attempt (29.9%) reattempted or died by suicide, with RR 2.92 (95% CI, 2.08 to 4.10). Among 46 patients without alcohol use disorder but with more than 2 suicide attempts in the past 3 years, 12 reattempted, with RR 2.11 (95% CI, 1.25 to 3.54). In patients without alcohol use disorder who had 1 or 2 previous suicide attempts in the past 3 years, anxiety disorder was associated with a subgroup in which 42 of 325 reattempted, although the reported RR was 0.98 (95% CI, 0.69 to 1.39).

Acute alcohol use during the index suicide attempt sharply separated higher-risk from lower-risk patients with alcohol use disorder. Among 127 patients with alcohol use disorder and acute alcohol use, 38 (29.9%) reattempted suicide or died by suicide within 6 months, with a relative event rate of 2.2 and RR 2.92 (95% CI, 2.08 to 4.10). By contrast, among 19 patients with alcohol use disorder but without acute alcohol use during the last attempt, none reattempted suicide during follow-up.

The lowest-risk subgroup was patients without alcohol use disorder, with fewer than 2 previous suicide attempts in the past 3 years, and without anxiety disorder. This group included 361 individuals, and 23 reattempted suicide during 6 months of follow-up. Compared with the root node, the RR was 0.36 (95% CI, 0.23 to 0.55), with a relative event rate of 0.49.

Among the 880 participants whose 6-month status was known, 117 patients (13.3%) repeated a suicide attempt and 11 (1.1%) died by suicide. In the complete-case analysis used for the survival tree, 115 of 878 patients (13.1%) had repeated suicide attempts.

The simplified algorithm divided patients into good-prognosis and poor-prognosis groups. The poor-prognosis group included patients with alcohol use disorder and acute alcohol use during the last suicide attempt, patients with anxiety disorder, and patients with a history of more than 2 suicide attempts in the past 3 years. All other patients were classified in the good-prognosis group.

The simplified algorithm had an AUC of 0.63 and identified a poor-prognosis group associated with higher risk of suicide reattempt, with HR 1.78 (95% CI, 1.42 to 2.24; P < .01). Its sensitivity was 0.8 (95% CI, 0.73 to 0.87), specificity was 0.47 (95% CI, 0.42 to 0.50), positive predictive value was 0.19 (95% CI, 0.15 to 0.22), and negative predictive value was 0.94 (95% CI, 0.92 to 0.96).

The authors used a survival tree because it can detect interactions among multiple clinical variables and sort patients into homogeneous subgroups based on time to suicide reattempt. Unlike a standard Cox model, which requires interactions to be specified in advance, the survival tree used recursive partitioning to identify the most relevant combinations of factors from the available baseline data. The authors presented it as a simple decision-support approach that may be easier for clinicians to use in practice.

The study analyzed 972 adults from the French ALGOS cohort who had survived a suicide attempt in the 7 days before inclusion. Participants were recruited from 23 French emergency services between January 2010 and February 2013. Patients were excluded if they had no suicidal intent, were homeless, were under guardianship, or were multirepeaters defined as more than 4 suicide attempts in the last 3 years.

The main limitations were 9.5% loss to follow-up, limited psychiatric assessment at baseline, and possible limits to generalizability outside France. Ninety-two of 972 participants were lost to follow-up at 6 months. Baseline assessment covered only 4 psychiatric disorders from the MINI, and the authors noted that the high prevalence of alcohol use disorder in France may limit extrapolation of the results to other settings.