3.1 Sample Characteristics
Sample characteristics are presented in Table 1. The mean age of participants was 72.6 (SD = 5.36), with a range of 60-92 years. Of the 6183 participants, 55.3% were female, 51.8% lived in the rural. Most participants were married (65.0%) and were not highly educated. In terms of participants’ type of living, 43.9% of them lived alone, and 31.0% lived with children. A total of 29.5% of the participants had chronic diseases. 58.5% never smoked, 47.5% never drank and 35.4% had “rarely” physical activity level. Among the respondents, 59.9% reported that they actively obtained health information. 27.5% reported getting health information from family and friends, while 33.5% reported getting information via Internet. The mean (SD) scores of eHEALS, GDS, PCS and MCS were 21.17 (8.25), 10.84 (6.30), 62.95 (15.83) and 63.91 (15.58), respectively. The eHEALS varied by gender, residence, marital status, education, employment status, type of living, stable income, chronic disease, smoking, drinking, physical activity, actively obtain health information, ways to seek health information and depression symptoms.
Table 1 General characteristics of the participants according to eHealth literacy (n = 6183)
Variables
|
n
|
%
|
eHealth Literacy
Mean (SD)
|
F / t-value
|
P-value
|
Sex
|
|
|
|
|
|
Female
|
3422
|
55.3
|
20.77 (8.38)
|
13.433
|
0.000
|
Male
|
2761
|
44.7
|
21.67 (8.07)
|
|
|
Residence
|
|
|
|
|
|
Rural
|
3205
|
51.8
|
19.57 (8.39)
|
108.69
|
0.000
|
City
|
2978
|
48.2
|
22.88 (7.74)
|
|
|
Marital Status
|
|
|
|
|
|
Unmarried
|
705
|
11.4
|
23.94 (8.09)
|
58.674
|
0.000
|
Married
|
4022
|
65.1
|
21.14 (8.08)
|
|
|
Divorce, bereaved, separated
|
1456
|
23.5
|
19.88 (8.47)
|
|
|
Education
|
|
|
|
|
|
No education or elementary school
|
2756
|
44.6
|
17.92 (8.41)
|
357.292
|
0.000
|
Middle school
|
1085
|
17.5
|
21.53 (7.07)
|
|
|
High school
|
882
|
14.3
|
23.80 (6.50)
|
|
|
College or higher
|
1460
|
23.6
|
25.46 (7.05)
|
|
|
Employment status
|
|
|
|
|
|
Unemployment
|
2392
|
38.7
|
21.17 (7.90)
|
179.594
|
0.000
|
Employment
|
865
|
14.0
|
24.95 (6.63)
|
|
|
Farm work
|
1800
|
29.1
|
18.10 (8.41)
|
|
|
Other
|
1126
|
18.2
|
23.16 (8.05)
|
|
|
Living alone
|
|
|
|
|
|
Yes
|
2714
|
43.9
|
20.85 (8.42)
|
-3.396
|
0.000
|
No
|
3469
|
56.1
|
21.57 (8.00)
|
|
|
Type of living- living with children
|
|
|
|
|
|
Yes
|
1915
|
31.0
|
21.27 (8.25)
|
9.163
|
0.002
|
No
|
1554
|
25.1
|
20.35 (8.26)
|
|
|
Religious belief
|
|
|
|
|
|
Yes
|
875
|
14.2
|
20.30 (8.68)
|
15.706
|
0.000
|
No
|
5308
|
85.8
|
21.31 (8.17)
|
|
|
Stable income
|
|
|
|
|
|
Yes
|
5122
|
82.8
|
23.18 (7.64)
|
92.650
|
0.000
|
No
|
1061
|
17.2
|
20.08 (8.37)
|
|
|
Chronic disease
|
|
|
|
|
|
Yes
|
2175
|
35.2
|
23.18 (7.64)
|
92.650
|
0.000
|
No
|
4008
|
64.8
|
20.08 (7.64)
|
|
|
Smoking
|
|
|
|
|
|
Never
|
3617
|
58.5
|
20.95 (8.47)
|
9.518
|
0.000
|
Rarely
|
1122
|
18.1
|
22.47 (7.57)
|
|
|
Sometimes
|
544
|
8.8
|
20.99 (7.30)
|
|
|
Often
|
589
|
9.5
|
20.26 (8.85)
|
|
|
Quit
|
311
|
5.0
|
21.11 (8.03)
|
|
|
Drinking
|
|
|
|
|
|
Never
|
2935
|
47.5
|
20.66 (8.60)
|
17.294
|
0.000
|
Rarely
|
1927
|
31.2
|
22.03 (7.81)
|
|
|
Sometimes
|
851
|
13.8
|
21.98 (7.41)
|
|
|
Often
|
364
|
5.9
|
19.78 (8.59)
|
|
|
Always
|
106
|
1.7
|
17.80 (9.00)
|
|
|
Physical activity
|
|
|
|
|
|
Never
|
985
|
15.9
|
18.23 (8.47)
|
48.441
|
0.000
|
Rarely
|
2186
|
35.4
|
20.95 (8.10)
|
|
|
Sometimes
|
1417
|
22.9
|
21.95 (7.67)
|
|
|
Often
|
1139
|
18.4
|
22.77 (7.99)
|
|
|
Always
|
456
|
7.4
|
21.17 (9.17)
|
|
|
Actively obtain health information
|
|
|
|
|
|
No
|
2478
|
40.1
|
18.07 (8.33)
|
142.388
|
0.000
|
Yes
|
3705
|
59.9
|
23.24 (7.51)
|
|
|
Ways to seek health information
|
|
|
|
|
|
Books and newspapers
|
1151
|
18.6
|
23.56 (7.26)
|
114.91
|
0.000
|
Family and friends
|
1699
|
27.5
|
19.25 (8.31)
|
|
|
Internet
|
2074
|
33.5
|
22.69 (7.77)
|
|
|
Medical staff
|
568
|
9.2
|
21.47 (8.47)
|
|
|
other
|
691
|
11.2
|
17.09 (8.18)
|
|
|
GDS-30
|
|
|
|
|
|
0~10
|
3019
|
48.8
|
22.01 (8.59)
|
55.082
|
0.000
|
11~20
|
2819
|
45.6
|
20.71 (7.73)
|
|
|
21~30
|
345
|
5.6
|
17.50 (8.12)
|
|
|
3.2 Psychometric properties of eHealth literacy
As shown in Table 2, the KMO value of sampling adequacy in the overall scale was 0.954. The Bartlett Test of Sphericity reached statistical significance (P = 0.000), which supported the factorability of the correlation matrix. Overall, one factor extracted from the 8 items of the eHEALS accounting for 78.52% of the variance. The factor loading values of 8 items were ranked from 0.832 to 0.905. The correlations between each item and the scale range from 0.835 to 0.904, indicating satisfactory convergent validity. Furthermore, reliability analyses showed good internal consistency in the eHEALS. The value of overall scale Cronbach α was 0.961.
Table 2 Construct, convergent, and criterion validity, and internal consistency of the 8-item eHealth Literacy Scale (n = 6183).
Construct validity, factor loadings
|
Values
|
eHealth Literacy scale items
|
|
I know what health resources are available on the internet
|
0.832
|
I know where to find helpful health resources on the internet
|
0.893
|
I know how to find helpful health resources on the internet
|
0.905
|
I know how to use the internet to answer my questions about health
|
0.902
|
I know how to use the health information I find on the internet to help me
|
0.902
|
I have the skills I need to evaluate the health resources I find on the internet
|
0.895
|
I have the skills needed to tell high-quality health resources from low-quality health resources on the internet
|
0.879
|
I feel confident in using information from the internet to make health decisions
|
0.880
|
Percentage of variance, %
|
78.52
|
Item-scale convergent validity, mean of Pa (range)
|
0.886 (0. 835~0.904)
|
Internal consistency, Cronbach α
|
0.961
|
a P: Spearman correlation coefficient.
3.3 Determinants of eHealth literacy
Results of the multivariate analysis depicted in Table 3 indicated that older adults with higher eHEALS scores were men (unstandardized regression coefficient [B] = 0.616, 95%CI 0.235~0.997, P = 0.002), those with living in the city (B = 1.471, 95%CI 1.064~1.878, P < 0.000), those with graduating from college or higher (B = 1.713, 95%CI 1.581~1.844, P < 0.000), and those with living with family (B = 6.436, 95%CI 4.574~8.298, P < 0.000), as compared to their counterparts, respectively. According to the results a multivariate analysis (Table 4), both chronic diseases (B = -2.218, 95%CI -2.632~-1.804, P < 0.000) and depression symptoms (B = -0.116, 95%CI 0.153~0.078, P < 0.000) were found negatively associated with eHEALS, while actively obtaining health information positively influenced eHEALS (B = 3.350, 95%CI 2.950~3.749, P < 0.000) after adjusting socio-demographics, drinking and SF-36.
Table 3 Multivariate line regression analyses for predictive factors of eHealth literacy
Variables
|
Unstandardized
Coefficients
|
Standardized
Coefficients
|
95%CI
|
t
|
P
|
|
B
|
β
|
|
|
|
Constant
|
5.604
|
|
3.239~7.969
|
4.646
|
0.000
|
Sex
|
0.616
|
0.037
|
0.235~0.997
|
3.170
|
0.002
|
Residence
|
1.471
|
0.089
|
1.064~1.878
|
7.078
|
0.000
|
Marital Status
|
-0.693
|
-0.049
|
-1.050~-0.335
|
-3.799
|
0.000
|
Education
|
1.713
|
0.067
|
1.581~1.844
|
25.489
|
0.000
|
Employment status
|
-0.146
|
0.053
|
-0.251~-0.041
|
2.732
|
0.006
|
Living alone
|
6.436
|
0.387
|
4.574~8.298
|
6.777
|
0.000
|
Living with children
|
1.472
|
0.321
|
0.963~1.981
|
5.667
|
0.000
|
Income
|
0.967
|
0.056
|
0.536~1.399
|
4.394
|
0.000
|
Table 4 Determinants of eHealth literacy among older adults
Variables
|
Model 1
|
Model 2
|
Model 3
|
|
B
|
95%CI
|
P value
|
B
|
95%CI
|
P value
|
B
|
95%CI
|
P value
|
Chronic diseases
|
-3.130
|
-3.547~-2.712
|
0.000
|
-2.446
|
-2.823~-2.011
|
0.000
|
-2.218
|
-2.632~-1.804
|
0.000
|
Physical activity
|
0.526
|
0.355~0.697
|
0.000
|
0.259
|
0.093~0.425
|
0.002
|
0.249
|
0.083~0.416
|
0.003
|
Actively obtain health information
|
4.468
|
4.063~4.873
|
0.000
|
3.438
|
3.041~3.834
|
0.000
|
3.350
|
2.950~3.749
|
0.000
|
Ways to seek health information
|
-0.369
|
-0.529~-0.264
|
0.000
|
-0.302
|
0.920~1.087
|
0.000
|
-0.309
|
0.918~1.089
|
0.000
|
Depression
|
-0.131
|
-0.169~0.092
|
0.000
|
-0.104
|
-0.141~-0.067
|
0.000
|
-0.116
|
-0.153~-0.078
|
0.000
|
MCS
|
-0.066
|
-0.093~-0.039
|
0.000
|
-0.033
|
-0.052~-0.013
|
0.000
|
-0.036
|
-0.054~-0.019
|
0.000
|
GH
|
0.031
|
0.016~0.046
|
0.000
|
0.036
|
0.021~0.050
|
0.000
|
0.039
|
0.024~0.053
|
0.000
|
Model 1: Adjusted smoking, drinking, PF, RP, BP, VT, SF, RE, MH, PCS
Model 2: Additionally, adjusted sex, residence, marital status, education level, employment status, religious belief
Model 3: Additionally, adjusted type of living, income, medical insurance
3.4 Mediation of eHealth literacy between depression and HRQoL
Figure 1 depicts how eHealth literacy mediates the relationship between depression symptoms and SF-36. Depression symptoms, an independent variable, was significantly associated with MCS and the dependent variable (B = -1.323, P < 0.000). Depression symptoms was also significant associated with eHealth literacy, a mediator (B = -0.220, P < 0.000). Further analysis of the indirect effects of eHealth literacy was conducted based on relationship between the main variables; when regressing MCS on depression symptoms and eHealth literacy, the B value decreased from -1.323 to -1.315 (P < 0.000). Moreover, depression symptoms, an independent variable, was significantly associated with PCS and the dependent variable (B = -1.111, P < 0.000) (Figure 2). Further analysis of the indirect effects of eHealth literacy was conducted based on relationship between the main variables; when regressing PCS on depression on eHealth literacy and depression symptoms, the B value decreased from -1.111 to -1.069 (P < 0.000). This implies that eHealth literacy partially mediated the relationship between depression symptoms and SF-36.