Hospitalisations at the end of life: using a sentinel surveillance network to study hospital use and associated patient, disease and healthcare factors

Table 3 Logistic regression analyses of factors associated with hospital use in the last three months of life (n = 319) *

		Beta	Standard Error	Odds-ratios [95%CI] † Hospitalised or not
Place of death wish expressed to GP	not expressed:			1.00 [1.00-1.00]
	home or living with family:	-0.603	0.348	0.547 [0.277–1.083]
	home for the elderly:	-1.404	0.493	0.246 [0.093–0.645]
	elsewhere (hospital, palliative care unit):	0.263	0.678	1.301 [0.345–4.908]
Palliative care delivered by GP	no			1.00 [1.00-1.00]
	yes	-1.405	0.366	0.245 [0.120–0.503]
Cause of death (ICD-10)	malignancies			1.00 [1.00-1.00]
	cardiovascular diseases	-1.305	0.390	0.271 [0.126–0.582]
	diseases of the nervous system	-0.284	0.902	0.753 [0.129–4.411]
	other	-1.053	0.374	0.349 [0.168–0.726]
Treatment goal in last three months before death ‡	palliative:			1.00 [1.00-1.00]
	curative/life-prolonging:	1.686	0.517	5.398 [1.958–14.882]
	from curative/life-prolonging to palliative:	0.830	0.353	2.293 [1.147–4.584]

* Multivariate logistic regression analyses for hospitalised (at least one day) (1) versus not hospitalised (0) patients; Missing values were included as a separate category if N>20; Variables entered in the regression were all variables significant in the bivariate analyses and cause of death, hospital beds in the healthcare area, degree of urbanization, and informal care provided [entered on the basis of results of other research e.g. (3)]. Interaction effects of the remaining variables were entered in the regression but not significant. No problems of multi-collinearity were observed.
† Odds-ratios with 95% Confidence Intervals are displayed; significant relationships are bold; Model summary results: Nagelkerke R Square = 0.368; Percentage correctly predicted = 76%;
‡ the subcategory "from palliative to curative/life-prolonging (n = 2) was not included in analyses
Abbreviations: GP = General Practitioners, ICD-10 = International Classification of Diseases version 10

ISSN: 1472-6963