Sedentary behaviour among older adults residing in flat and hilly neighbourhoods and its association with frailty and chronic disease status

Table 4 The associations of sedentary behaviour with health indicators and covariates between flat and hilly neighbourhoods

Predictor variable	Flat neighbourhood (n = 704)				Hilly neighbourhood (n = 505)
	Coefficients				Coefficients
	B	SE	β(t)	95% CI	B	SE	β(t)	95% CI
Baseline models
(Constant)	90.567	3.952	(22.92)**	± 15.518	-1213.56	489.596	(-2.479)*	± 1924.585
Frailty	-1.823	0.597	-0.156(-3.05)*	± 2.345	95.037	22.954	0.195(4.14)**	± 90.23
CDS	11.738	3.957	0.118(2.967)*	± 15.536	461.04	110.114	0.212(4.19)**	± 432.855
Poor health	3.382	4.999	0.034(0.676)	± 19.629	-90.347	126.334	-0.036(-0.72)	± 496.615
Age-adjusted models
(Constant)	111.616	12.528	(8.91)**	± 49.206	-2634.05	510.185	(-5.16)**	± 2005.535
Frailty	-1.477	0.658	-0.122(-2.246)*	± 2.583	69.355	22.155	0.143(3.13)*	± 87.09
CDS	14.228	4.456	0.146(3.193)*	± 17.501	296.683	107.472	0.136(2.76)*	± 422.472
Poor health	-0.891	5.537	-0.009(-0.161)	± 21.748	-87.519	120.159	-0.035(-0.73)	± 472.345
Age (yrs)	-0.365	0.213	-0.077(-1.718)	± 0.835	33.35	4.894	0.314(6.82)**	± 19.238
Ultimate (fully adjusted) models
(Constant)	84.023	12.846	(6.541)**	± 50.469	-2477.06	519.367	(-4.77)**	± 2041.667
Frailty	-2.082	0.651	-0.182(-3.20)**	± 2.559	78.062	22.044	0.160(3.54)**	± 86.656
CDS^a	10.71	4.38	0.115(2.45)*	± 17.206	281.867	106.423	0.130(2.65)*	± 418.355
Poor health^b	-1.67	5.48	-0.017(-0.31)	± 21.529	-38.331	119.997	-0.015(-0.32)	± 471.715
Age (yrs)	0.16	0.22	0.034(0.73)	± 0.864	33.985	5.062	0.320(6.71)**	± 19.901
Women^c	20.022	3.54	0.235(5.66)**	± 13.976	-337.246	101.973	-0.152(-3.31)**	± 400.865
Income (₵)	0.00	0.003	0.006(0.14)	± 0.013	-0.104	0.052	-0.091(-2.02)*	± 0.203

SE Standard error (of B), CI Confidence interval (of B), CDS Chronic disease status (i.e., one or more conditions), B Unstandardised coefficient, β Standardised coefficient; sedentary behaviour is the dependent variable in all models; F-test was significant at a minimum of p < 0.05 for all models; the total variance explained by predictors in the flat neighbourhood models ranged from 2.1 to 8.2% whereas the variance explained in the hilly neighbourhood models ranged from 9.2 to 20.2%; tolerance ≥ 0.5 for all predictors in the flat and hilly neighbourhood models, and Durbin-Watson values for the models were approximately 2
^a ‘None’ set as a reference group
^b ‘Good health’ set as a reference group
^c ‘Men’ set as a reference group
**p < 0.001; *p < 0.05

ISSN: 1471-2458