# Prediction of suicide attempts in depressive disorder and panic disorder with/without agoraphobia: multivariate analysis of clinical predictors

## TL;DR

This study identifies suicidal ideation and depression as key predictors of suicide attempts in patients with depressive disorder, while panic and agoraphobic symptoms may reduce risk.

## Contribution

The study introduces a predictive model for suicide attempts in depressive and panic disorders, revealing suppressor effects of panic/agoraphobic symptoms.

## Key findings

- Suicidal ideation and depression severity were the strongest predictors of suicide attempts in depressive disorder.
- Panic and agoraphobic symptoms showed a suppressor effect, reducing suicide risk by 11%.
- The predictive model failed to identify the single suicide attempt in the panic disorder group, showing limited utility there.

## Abstract

Suicide risk in patients with depressive disorder (DD) and panic disorder with or without agoraphobia (PD/A) is often associated with depression, hopelessness, and anxiety-related factors, but their predictive roles remain unclear.

To identify the strongest clinical predictors of suicide attempts in patients with DD and PD/A, and to assess the applicability of a predictive model within and across diagnostic groups.

A total of 94 patients (48 with DD- acute or recurrent, 46 with PD/A) were assessed using the Beck Scale for Suicide Ideation (BSSI), Beck Depression Inventory-II (BDI-II), Beck Hopelessness Scale (BHS), Anxiety Sensitivity Index-Revised (ASI-R), and Beck Anxiety Inventory (BAI). Univariate and multivariate analyses were conducted to evaluate the relative predictive value of these factors for suicide attempts within and across diagnostic groups.

Suicide attempts were reported by 23 patients with DD (48%) and only one patient with PD/A (2.17%). In DD sample, suicidal ideation was the strongest predictor of suicide attempts, followed by depression severity. ROC analyses indicated optimal cut-offs of 11 for BSSI and 35 for BDI-II, yielding high sensitivity (82.6%) and specificity (92%). Hopelessness and anxiety sensitivity did not retain predictive value in multivariate analysis, whereas panic/agoraphobic symptoms showed a suppressor effect, reducing risk by 11%. When applied to the PD/A group, the model estimated an average predicted risk of 9.3%, but failed to identify the single attempter, resulting in zero sensitivity.

High levels of suicidal ideation and depressive symptoms were the strongest predictors of suicide attempts in DD. Panic and agoraphobic symptoms exerted a suppressor effect on suicide risk. In the PD/A group, however, the model’s predictive utility was extremely limited due to the very low number of suicide attempts, and these findings should therefore be considered preliminary.