Table 1.

Characteristics associated with recognising/not recognising death in the near future in patients who died non-suddenly at home/in a care home, n = 252^a

				Logistic regressionb (odds ratio [95%CI])
Patient and care characteristics	Total %	Did not recognise death in the near future (n = 72),%	Recognised death in the near future (n = 175),%	Univariate analyses	Multivariate analysesc
Age, years
1–64	20	18	21	1
65–85	41	42	40	0.84 (0.4 to 1.8)	d
≥85	39	39	39	0.86 (0.4 to 1.9)	d
Sex
Male	45	53	42	1
Female	55	47	58	1.53 (0.9 to 2.6)	d
Education
Elementary	48	59	44	1
Secondary	39	31	41	1.77 (0.9 to 2.6)	d
Tertiary	13	9	14	2.02 (0.8 to 5.4)	d
Dutch nationality
≥1 parent	94	96	93	1
≤1 parent	6	4	7	1.58 (0.4 to 5.8)	d
Cause of death
Cancer	38	15	47	1	1
Not cancer	62	85	53	0.20 (0.1 to 0.4)	0.18 (0.1 to 0.4)
Dementia diagnosed by a physician
No	88	88	88	1
Yes	12	12	12	0.95 (0.4 to 2.3)	d
Related symptoms and functional state before death
Comatose
No	48	59	48	1
Yes	52	41	52	0.64 (0.3 to 1.2)	d
Lack of appetite
No	22	46	17	1
Yes	78	54	83	4.10 (1.9 to 8.8)	d
Lack of energy
No	8	18	6	1
Yes	92	82	94	3.68 (1.3 to 10.6)	d
Pain
No	52	58	50	1
Yes	48	42	50	1.39 (0.7 to 2.8)	d
Difficulty breathing
No	59	66	57	1
Yes	41	34	43	1.45 (0.7 to 3.0)	d
Anxiety
No	60	58	60	1
Yes	40	42	40	0.94 (0.4 to 2.0)	d
Low functional status capable of only limited self-care (ECOG score ≥4)e
No	85	32	8	1	1
Yes	15	68	92	5.39 (2.5 to 11.5)	5.21 (2.3 to 11.7)

ECOG = Eastern Cooperative Oncology Group.

^aIncludes five missing values; percentages of missing observations variables ranged between 0.4% and 5.6%.

^bDependent variable: patients in whom GPs ever recognised death in the near future, n = 175; reference group: patients who died without their GPs recognising death in the near future, n = 72.

^cStepwise backwards logistic regression. Variables removed after three steps of the backward analyses. Significant values are in bold.

^dNot entered/retained following multiple backwards logistic regression analyses.

^eECOG scale.²⁹