The association between self-esteem and suicidal risk: a meta-analysis

• Published in: European psychiatry Vol. 65; no. S1; p. S835
• Main Authors: Lippo, F., Madeddu, F., Fornaro, M., Calati, R.
• Format: Journal Article
• Language: English
• Published: Paris Cambridge University Press 01-06-2022
• Subjects: Abstract; E-Poster Viewing; Meta-analysis; Self esteem; suicidal risk; Suicide; Suicides & suicide attempts
• ISSN: 0924-9338; 1778-3585
• DOI: 10.1192/j.eurpsy.2022.2162

Introduction
Background: Existing evidence poses low self-esteem as a risk factor for both suicidal ideation (SI) and suicide attempts (SAs) in the general population.
Objectives
The present study assesses the relationship between self-esteem level and SI/SA, considering across the lifespan. Two separate meta-analyses, one for SI and the other for SA are herein reported since they substantially overlap in terms of eligibility procedures and search strategies.
Methods
Eligible studies documented at least one suicidal, and a non-suicidal group. Data were analyzed using the Cochrane Collaboration Review Manager Software (RevMan, version 5.4.1) under the random-effects models. Values were standardized owing to the anticipated heterogeneity of self-esteem rating tools. Sensitivity analyses were performed to control for heterogeneity.
Results
Out of 3,310 initial hits, 24 studies were deemed eligible for inclusion. The meta-analyses showed that individuals with lower levels of self-esteem, compared to those with higher levels, were more likely to endorse both SI and SA. SI reached a standardized mean difference of -0.43 (CI: -0.81, -0.05), while SA reduced by -0.89 (CI: -1.02, -0.76), overall. Limitations: The herein presented results rely on standardized mean differences rather than odds of either SI or SA since the original studies failed to systematically fetch rates of the events.
Conclusions
Lower levels of self-esteem represent a risk factor for both SI and SA across the lifespan. Forthcoming studies should systematically account for multiple moderators to allow meta-analytic synthesis including sub-group and meta-regression analyses assuming high-heterogeneity would still be concerned.
Disclosure
No significant relationships.