Association between living arrangements and cognitive decline in older adults: A nationally representative longitudinal study in China

Table 3 Stratification analysis of the association between living arrangements and cognitive decline

	Males (n = 3016)	Females (n = 3058)
Fixed Effects
Intercept	4.634^***	6.000^***
Living arrangement (ref = Living with spouse)
Living alone	0.004	0.159
Living with adult children	-0.037	-0.126
Living with spouse and adult children	0.065	-0.084
Living with others	-0.019	0.182
Time (years since baseline)	-0.021	-0.077^***
Age (ref = 60–64, at baseline)
65–69	-0.293^**	-0.673^***
70–74	-1.034^***	-1.165^***
≥ 75	-1.818^***	-1.940^***
Geographic residence (ref = Rural)
Urban	0.236^*	0.591^***
Education (ref = No formal education)
Capable of reading and/or writing	2.341^***	2.274^***
Primary school	3.660^***	4.179^***
Middle school and above	4.800^***	5.398^***
Working status (ref = No)
Yes	0.060	-0.081
The number of physical comorbidities (ref = 0)
1–2	0.146	-0.066
≥ 3	0.182	0.156
Feeling pain (ref = No)
Yes	-0.283^*	-0.103
IADLs (ref = Unimpaired)
Impaired	-0.837^***	-0.782^***
Depressive symptoms (ref = No)
Yes	-0.460^***	-0.477^***
Social activity participation (ref = No)
Yes	0.656^***	0.510^***
Number of adult children (ref = 1)
2–3	0.249	0.022
≥ 4	0.231	-0.089
Average schooling year of children (Centered)	0.174^***	0.237^***
Household expenditure per capita (log)	0.360^***	0.154^**
Time * Living alone	-0.192^***	-0.071
Time * Living with adult children	-0.168^**	-0.107^**
Time * Living with spouse and adult children	-0.032	-0.053^*
Time * Living with others	-0.061	-0.179^*
Random Effects
Level 2: Individual
Individual -variance	4.011^***	5.274^***
Level 1: Point in time
Point in time-variance	6.795^***	6.135^***
*-2loglikelihood**	49282.55	48645.26

Note: (1) Figures in the table were parameter estimates based on unweighted multilevel models from 3106 males consisting of 9699 observations and 3058 females consisting of 9606 observations. (2) Model estimation was based on the iterative generalized least squares. (3) ref = the reference category, IADLs = instrumental activity of daily living. (4) ^*P < 0.05, ^**P < 0.01, ^***P < 0.001

ISSN: 1471-2318