Table 4.

Regression Analysis with Social Isolation as the Dependent Variable (Wave 4), and Frailty, Social Isolation, and Covariate/Confounds (Wave 2) Used as Predictor Variables.

	B	β	T	p	Lower Bound CI (95%)	Higher Bound CI (95%)	B	β	t	p=	Lower Bound CI (95%)	Higher Bound CI (95%)
	Frailty index (n = 1457)						Frailty phenotype (n = 1131)
	Sample 4						Sample 5
Step 1
Age	.01	.02	.68	.494	−.010	.020	.01	.03	1.10	.274	−.007	.026
Gender	−.18	−.04	−1.84	.067	−.362	.012	−.23	−.06	−2.18	.029	−.440	-.023
Social isolation from a wider social network (Wave 2)	.50	.48	21.02	.000	.449	.541	.54	.52	20.76	.000	.490	.592
Educational qualification	−.11	−.12	−5.11	.000	−.155	−.069	−.10	−.11	−4.14	.000	−.147	-.052
Smoking	.01	.01	.05	.961	−.182	.191	.10	.02	.96	.339	−.105	.306
Depressive symptoms	.01	.01	.03	.975	−.094	.097	.02	.01	.34	.736	−.092	.130
Psychiatric symptoms	.05	.01	.15	.883	−.591	.687	.24	.01	.52	.601	−.653	1.129
Wealth	−.01	−.05	−2.11	.035	.000	.000	−.01	−.06	−2.40	.016	.000	.000
Loneliness	.07	.05	2.04	.041	.003	.136	.03	.02	.69	.493	−.049	.102
Step 2
Frailty (Wave 2)	.18	.06	2.51	.012	.040	.328	.20	.06	2.28	.023	.028	.372

Abbreviations: B = unstandardised beta; β = standardised betas; t = t test value; p = probability; CI = confidence interval.

Note. Social isolation from a wider social network at Wave 4 is used as a dependent variable, and gender, age, social isolation from a wider social network at Wave 2, education level, whether smoker during lifetime, depressive symptoms, psychiatric symptoms in the last two years, wealth, and loneliness (Step 1), and frailty at Wave 2 (frailty index/Fried phenotype) (Step 2) are used as predictor variables.