. 2020 Apr 21;30(4):539–548. doi: 10.1007/s00787-020-01534-3

Table 2.

Odds ratios (95% confidence intervals) from the univariate and multivariate conditional logistic regression analyses indicating risk of hospital-treated deliberate self-harm associated with the variables under study, for all subjects and males and females separately. Log-likelihood ratio tests for gender differences

Variable	Crude OR^a	Adjusted OR^b			Interaction with gender^c
Variable	All subjects	All subjects	Females	Males	χ²	p
Bereavement status					1.30	0.25
No exposure to parental DBEC	1 (reference)	1 (reference)	1 (reference)	1 (reference)
Exposure to parental DBEC	2.49 (2.20–2.83)^‡	1.83 (1.60–2.10)^‡	1.98 (1.66–2.36)^‡	1.69 (1.37–2.08)^‡
Cause of death					4.31	0.37
No exposure to parental DBEC	1 (reference)	1 (reference)	1 (reference)	1 (reference)
Suicide	3.07 (2.58–3.66)^‡	2.32 (1.92–2.80)^‡	2.54 (1.99–3.24)^‡	2.06 (1.53–2.78)^‡
Transport accidents	1.62 (1.23–2.15)^‡	1.26 (0.93–1.69)	1.12 (0.72–1.73)	1.42 (0.94–2.15)
Other accidents	2.69 (2.11–3.42)^‡	1.79 (1.38–2.33)^‡	1.97 (1.42–2.75)^‡	1.63 (1.07–2.49)*
Other external causes	1.66 (0.84–3.29)	1.13 (0.55–2.31)	1.72 (0.73–4.05)	0.54 (0.13–2.28)
Gender of deceased					2.16	0.54
No exposure to parental DBEC	1 (reference)	1 (reference)	1 (reference)	1 (reference)
Father	2.38 (2.06–2.74)^‡	1.76 (1.51–2.06)^‡	1.85 (1.51–2.26)^‡	1.71 (1.35–2.16)^‡
Mother	2.78 (2.13–3.63)^‡	2.00 (1.51–2.66)^‡	2.37 (1.64–3.42)^‡	1.58 (1.00–2.48)
Both parents	5.12 (2.34–11.20)^‡	3.09 (1.30–7.36)*	3.19 (1.19–8.53)*	2.82 (0.50–15.79)
Age at bereavement					3.35	0.50
No exposure to parental DBEC	1 (reference)	1 (reference)	1 (reference)	1 (reference)
≤ 4 years	2.38 (1.84–3.07)^‡	1.69 (1.29–2.23)^‡	1.64 (1.12–2.42)*	1.80 (1.22–2.68)^†
5–9 years	2.46 (1.93–3.14)^‡	1.77 (1.37–2.30)^‡	2.09 (1.50–2.91)^‡	1.41 (0.92–2.16)
10–14 years	2.37 (1.88–2.98)^‡	1.81 (1.42–2.32)^‡	2.02 (1.48–2.76)^‡	1.53 (1.02–2.31)*
15–18 years	2.90 (2.22–3.78)^‡	2.14 (1.60–2.87)^‡	2.18 (1.47–3.23)^‡	2.15 (1.40–3.31)^‡

DBEC death by external causes

*p < 0.05, ^†p < 0.01, ^‡p < 0.001

^aThe ORs derived from these models were adjusted for age, gender, and calendar time through matching

^bThe ORs derived from these models were further adjusted for ethnicity, highest completed education, taxable income, marital status, marital stability, residence stability, number of sick leave pension episodes, and diagnosis of sick leave pension episode

^cInteractions with gender of bereaved offspring were based on multivariate analyses