Table 1.
Clinical risk factors for self-harm in prison
| Studies analysed (k) | Participants (n) | OR (95% CI) | z score | p value | Heterogeneity (I2) | ||
|---|---|---|---|---|---|---|---|
| Suicidal ideation | |||||||
| Current or recent | 6 | 7256 | 13·8 (8·6–22·1) | 10·9 | <0·0001 | 49% | |
| Lifetime history | 5 | 3779 | 8·9 (6·1–13·0) | 11·3 | <0·0001 | 56% | |
| Overall | 11 | 11 035 | 10·9 (8·0–14·9) | 15·1 | <0·0001 | 58% | |
| Psychiatric treatment | |||||||
| In prison | 7 | 30 931 | 10·5 (4·8–22·8) | 5·9 | <0·0001 | 93% | |
| Before prison | 14 | 11 001 | 3·7 (2·8–4·9) | 9·4 | <0·0001 | 58% | |
| Current psychiatric diagnosis | |||||||
| Any psychiatric disorder | 4 | 134 954 | 8·1 (7·0–9·4) | 27·6 | <0·0001 | 0% | |
| Major depression | 4 | 3908 | 9·3 (2·9–29·5) | 3·8 | <0·0001 | 91% | |
| Borderline personality disorder | 3 | 3932 | 9·2 (3·7–22·5) | 4·8 | <0·0001 | 81% | |
| Psychotic disorder | 5 | 4172 | 4·4 (2·5–7·7) | 5·0 | <0·0001 | 38% | |
| Anxiety disorder | 4 | 3908 | 4·2 (2·3–7·7) | 4·7 | <0·0001 | 74% | |
| Substance use disorder | 3 | 766 | 2·3 (1·4–3·9) | 3·1 | 0·0018 | 43% | |
| Antisocial personality disorder | 5 | 4172 | 1·0 (0·5–2·1) | 0·1 | 0·90 | 89% | |
| Previous self-harm | 19 | 14 154 | 6·6 (5·3–8·3) | 16·3 | <0·0001 | 55% | |
| Impulsivity | 5 | 2153 | 4·0 (2·6–6·3) | 6·2 | <0·0001 | 61% | |
| Hopelessness | 4 | 6201 | 3·9 (2·1–7·2) | 4·3 | <0·0001 | 53% | |
| Current psychotropic medication | 8 | 6400 | 3·6 (2·5–5·2) | 6·6 | <0·0001 | 77% | |
| Psychological distress | |||||||
| Severe (cutoff)* | 7 | 9162 | 3·4 (2·0–5·9) | 4·4 | <0·0001 | 85% | |
| Continuous | 6 | 6345 | 3·4 (2·1–5·4) | 5·1 | <0·0001 | 97% | |
| Physical health problems | 9 | 30 479 | 2·3 (1·9–2·8) | 9·0 | <0·0001 | 14% | |
| History of illicit drug use | 10 | 34 978 | 2·0 (1·3–3·0) | 3·2 | 0·0015 | 82% | |
| History of alcohol misuse | 8 | 13 773 | 1·4 (0·8–2·5) | 1·4 | 0·18 | 87% | |
I2 represents the percentage of variability in estimates of effect size that is attributable to between-study variation (heterogeneity). OR=odds ratio.
*
The cutoff value depended on the study and scale.