Table 3. The relationship between potential explanatory variables and lifetime severe suicidal ideation (high-risk patients): results from the binary logistic regression analysis (n=129).
| B | SE | Wald | p-value | OR | 95%CI | |
|---|---|---|---|---|---|---|
| Highest stressful life event | 0.141 | 0.050 | 8.033 | 0.005 | 1.151 | 1.044-1.269 |
| HAM-D | 0.183 | 0.055 | 11.187 | 0.001 | 1.201 | 1.079-1.337 |
| DUI > 24 months | 2.167 | 1.112 | 3.796 | 0.051 | 8.731 | 0.987-77.217 |
| Constant | -7.015 | 1.503 | 21.776 | < 0.001 | 0.001 | - |
χ2 = 28.842, p < 0.001, R2 Nagelkerke = 0.380
DUI = duration of untreated illness; HAM-D = Hamilton Depression Rating Scale; OR = odds ratios; SE = standard error.
Variable inflation factor (collinearity) between 1.022 and 1.047, indicating no collinearity. Total number of variables initially entered in the model n=13; variables not retained in the model: sex, number of life events, life events (≥ 1), severe life events (≥ 1), Hamilton Rating Scale for Anxiety (HAM-A), Beck Depression Inventory (BDI) lifetime, substance use disorder lifetime, duration of illness, family history of mood disorders, DUI.