Table 2.

Base Mixed Models for Caregiver Outcomes without Covariates

	Estimate	SE	T-statistic	p-value
Model 1: Affect Balance
Intercept	4.40	0.60	7.35	<0.01 *
Time	0.84	0.27	3.10	<0.01 *
Baseline Closeness	0.16	0.03	5.25	<0.01 *
Baseline Closeness*Time	−0.05	0.01	−3.34	<0.01 *
Model 2: Affect Balance
Intercept	7.55	0.13	57.76	<0.01 *
Time	−0.10	0.06	−1.79	0.08
Closeness Change	0.09	0.04	2.33	0.02 *
Change in Closeness* Time	−0.05	0.02	−2.65	<0.01 *
Model 3: Beck Depression Inventory
Intercept	15.04	2.81	5.35	<0.01 *
Time	−1.67	1.31	−1.27	0.20
Baseline Closeness	−0.35	0.15	−2.4	0.02 *
Baseline Closeness*Time	0.07	0.07	1.04	0.30
Model 4: Beck Depression Inventory
Intercept	8.49	0.63	13.51	<0.01 *
Time	−0.35	0.30	−1.17	0.24
Closeness Change	0.05	0.20	0.27	0.79
Closeness Change * Time	−0.03	0.10	−0.28	0.78
Model 5: SF-12: Mental Health Component
Intercept	38.00	2.59	14.67	<0.01 *
Time	3.16	1.40	2.24	0.02 *
Baseline Closeness	0.68	0.13	5.10	<0.01 *
Baseline Closeness*Time	−0.18	0.07	−2.51	0.01 *
Model 6: SF-12: Mental Health Component
Intercept	51.07	0.58	87.74	<0.01 *
Time	−0.53	0.30	−1.79	0.08
Closeness Change	0.29	0.19	1.57	0.11
Closeness Change * Time	−0.30	0.11	−2.80	<0.01 *
Model 7: SF-12: Physical Health Component
Intercept	48.13	3.32	14.49	<0.01 *
Time	−2.03	1.28	−1.59	0.11
Baseline Closeness	0.03	0.17	0.19	0.85
Baseline Closeness*Time	0.09	0.07	1.42	0.16
Model 8: SF-12: Physical Health Component
Intercept	49.00	0.71	68.55	<0.01 *
Time	−0.11	0.25	−0.44	0.66
Closeness Change	0.06	0.23	0.26	0.79
Closeness Change * Time	0.22	0.10	2.18	0.03 *

Note: Negative scores in Closeness Change indicate a greater loss of closeness from predementia levels to postdementia levels.

^*indicates statistical significance at a level of p < 0.05.