Skip to main content

Table 3 Self-harm behaviour: odds ratios and 95% confidence intervals in logistic regression

From: Could a brief assessment of negative emotions and self-esteem identify adolescents at current and future risk of self-harm in the community? A prospective cohort analysis

  Males: adjusted* OR (95% CI) Females: adjusted* OR (95% CI)
Baseline self-harm   
Personal failure 1.18 (1.08 to 1.30) 1.24 (1.15 to 1.34)
Generalised anxiety 1.06 (0.97 to 1.16) 1.02 (0.95 to 1.10)
Physical symptoms 1.25 (1.15 to 1.35) 1.16 (1.07 to 1.24)
Positive self-esteem 0.95 (0.87 to 1.04) 0.82 (0.75 to 0.89)
6 month self-harm   
Personal failure 1.22 (1.12 to1.32) 1.19 (1.11 to 1.28)
Generalised anxiety 1.03 (0.95 to 1.12) 1.03 (0.97 to 1.11)
Physical symptoms 1.12 (1.04 to 1.21) 1.14 (1.06 to 1.22)
Positive self-esteem 0.97 (0.90 to 1.06) 0.87 (0.81 to 0.94)
12 month self-harm   
Personal failure 1.05 (0.95 to 1.16) 1.16 (1.08 to 1.25)
Generalised anxiety 1.08 (1.00 to 1.18) 1.01 (0.95 to 1.08)
Physical symptoms 1.16 (1.07 to 1.25) 1.12 (1.05 to 1.2)
Positive self-esteem 1.01 (0.93 to 1.1) 0.88 (0.82 to 0.95)
Future self-harm models: further adjusted for baseline self-harm thoughts and acts
6 month self-harm   
Personal failure 1.13 (1.03 to 1.24) 1.06 (0.98 to 1.16)
Generalised anxiety 0.98 (0.9 to 1.08) 1.02 (0.95 to 1.1)
Physical symptoms 1.05 (0.96 to 1.14) 1.08 (0.99 to 1.16)
Positive self-esteem 1 (0.92 to 1.1) 0.91 (0.84 to 0.99)
Self-harm thoughts 4.12 (2.44 to 6.97) 3.36 (2.13 to 5.29)
Self-harm acts 3.58 (2.06 to 6.22) 5.8 (3.68 to 9.13)
12 month self-harm   
Personal failure 0.97 (0.87 to 1.07) 1.06 (0.98 to 1.14)
Generalised anxiety 1.05 (0.96 to 1.15) 0.99 (0.93 to 1.06)
Physical symptoms 1.1 (1.01 to 1.2) 1.06 (0.99 to 1.15)
Positive self-esteem 1.04 (0.94 to 1.13) 0.91 (0.84 to 0.98)
Self-harm thoughts 2.25 (1.29 to 3.92) 3.1 (2.03 to 4.75)
Self-harm acts 4.69 (2.65 to 8.3) 2.91 (1.86 to 4.55)
  1. *All models adjusted for age. NB: All scales within each model were entered simultaneously and therefore OR and CIs are for independent effects of each scale. Where there is strong evidence of association, this is highlighted in bold text.