Table 3.

Odds ratios for the association of potential confounders with non-fatal self harm and completed suicides

	Non-fatal self harm			Completed suicides
Variable	Cases (n=1968)	Controls (n=34 494)	Crude odds ratio (95% CI)*	Cases (n=69)	Controls (n=1121)	Crude odds ratio (95% CI)
Severity of depression:
Mild	1260	22095	1	54	785	1
Moderate	601	11047	0.98 (0.88 to 1.08)	13	293	0.64 (0.33 to 1.22)
Severe	107	1352	1.43 (1.16 to 1.76)	2	43	—
Referral to psychiatrist or psychologist†	202	891	4.10 (3.48 to 4.84)	6	26	3.41 (1.28 to 9.05)
History of self harm‡	64	134	9.14 (6.73 to 12.4)	3	4	15.0 (3.36 to 67.0)
Diagnosis of, or treatment for, anxiety or panic disorder	643	7888	1.67 (1.51 to 1.85)	19	228	1.60 (0.91 to 2.82)
Schizophrenia	2	8	1.82 (0.32 to 10.5)	1	3	—
Antipsychotic therapy	255	1908	2.40 (2.08 to 2.77)	12	46	4.67 (2.30 to 9.51)
Drug misuse	2	9	2.90 (0.61 to 13.8)	0	2	—
Alcohol misuse	213	1173	3.58 (3.04 to 4.21)	7	53	2.31 (0.97 to 5.49)
Lithium therapy§	23	182	1.81 (1.16 to 2.83)	0	6	—
No of different antidepressants prescribed§:
0	715	19 536	1	27	673	1
1	807	11 544	1.97 (1.77 to 2.19)	29	365	2.05 (1.18 to 3.59)
2 or 3	422	3232	3.65 (3.19 to 4.18)	12	82	3.42 (1.57 to 7.43)
>3	24	182	3.41 (2.18 to 5.33)	1	1	—

^*Values not shown for age and sex as they were matching variables.

^†In year before index day.

^‡Includes drug overdose, deliberate poisoning, self laceration, and other methods of self harm before entry to cohort.

^§In year before index day and since entry to cohort.