Skip to main content
. 2003 Jul 12;327(7406):81. doi: 10.1136/bmj.327.7406.81

Table 6.

Univariate and multivariate (after stepwise exclusion) Cox regression analysis of mortality, 1992-2002, among 185 homeless people interviewed in two Copenhagen hostels

Variable No (%) Unadjusted relative risk (95% CI) Adjusted relative risk (95% CI)
Mean (SD) age (years) 38.2 (10.9) 1.03* (1.00 to 1.05) 1.05*** (1.02 to 1.08)
Male 161 (87) 1.50 (0.60 to 3.26)
School education 1.01 (0.84 to 1.22)
Mother died before interviewee was 17 20 (11) 1.19 (0.54 to 2.62)
Father died before interviewee was 17 25 (14) 1.91* (1.01 to 3.62)
Seldom or no contact with family 81 (44) 1.02 (0.60 to 1.73)
Institutionalised as child 19 (10) 0.83 (0.33 to 2.09)
Misuse of alcohol 86 (47) 2.04** (1.19 to 3.50) 2.19** (1.25 to 3.86)
Misuse of drugs:
Opioids 81 (44) 0.91 (0.54 to 1.56)
Cannabis 48 (26) 0.89 (0.48 to 1.65)
Sedatives 25 (14) 2.02* (1.07 to 3.82) 3.14** (1.55 to 6.37)
Cocaine or stimulating drugs 25 (14) 1.60 (0.81 to 3.16))
Any misuse 144 (78) 1.37 (0.69 to 2.70)
Psychosis 60 (32) 0.64 (0.35 to 1.18) 0.47* (0.24 to 0.93)
Ever sentenced 135 (73) 1.01 (0.56 to 1.82)
Sentenced for violence 47 (25) 0.86 (0.46 to 1.60)
*

P<0.05

**

P<0.01

***

P<0.001.

Adjusted for other variables in final model.

Estimate of increased risk for increased level of school education: 7 years, 8 years, 9 years, 10 years, and A level.

HHS Vulnerability Disclosure