How to integrate proxy data from two informants in life event assessment in psychological autopsy

BMC Psychiatry

Table 1 Comparison of characteristics of 1st and 2nd informants of the target

	1st informant n (%)	2nd informant n (%)	χ²/t	df	P
Gender			25.71	1	< 0.001
Male	159 (38.2)	232 (55.8)
Female	257 (61.8)	184 (44.2)
Age (yr)	36 (29, 46) ^a	31 (21, 41) ^a	6.36^b	827	< 0.001
Marital status			51.43	3	< 0.001^c
Single	35 (8.4)	112 (26.9)
Married	371 (89.2)	298 (71.7)
Widowed	9 (2.2)	5 (1.2)
Others	1 (0.2)	1 (0.2)
Education (yr)	9 (6, 9) ^a	9 (7, 9) ^a	− 3.65^b	830	< 0.001
Religion			4.36	4	0.317^c
Atheism	378 (90.9)	386 (92.8)
Catholicism	10 (2.4)	8 (1.9)
Buddhism	26 (6.2)	20 (4.8)
Other religion	0 (0)	2 (0.5)
Data missing	2 (0.5)	0 (0)
Annual family income (1000 RMB)	14.3 (10.0, 25.0) ^a	15.0 (10.0, 25.0) ^a	− 0.68^b	728	0.498
CES-D score	4 (1, 9) ^a	4 (1, 9) ^a	0.29^b	826	0.771
Familiarity to the target			78.45	5	< 0.001^c
Very unfamiliar	0(0)	1(0.2)
Unfamiliar	7(1.7)	9(2.2)
Middle	67(16.1)	154(37.0)
Familiar	159(38.2)	155(37.3)
Very familiar	139(33.4)	51(12.2)
Data missing	44(10.6)	46(11.1)

^aBecause of non-normal distributions, median (1st, 3rd quartiles) was used. ^bt test was used for those numerical variables
^cFisher’s exact test was employed

ISSN: 1471-244X