Table 4.
The mediating effect of health literacy and health behaviors on family health and frailty explored by Sobel tests and Karlson-Holm-Breen (KHB) decomposition methods.
|
|
Health literacy | Smoking status | Alcohol intake | Physical exercise | Sleep duration | Breakfast | |||||||
| Family health→mediator, β | .286a | .006a | .001 | −.002 | .012a | .018a | |||||||
| Mediator→frailty, β | −.008a | −.105a | −.023 | −.138a | −.095a | −.154a | |||||||
| Indirect effect, β | −.002a | −.001a | −.000 | .000 | −.001a | −.003a | |||||||
| Direct effect, β | −.014a | −.016a | −.016a | −.017a | −.015a | −.014a | |||||||
| Total effect, β | −.016a | −.016a | −.016a | −.016a | −.016a | −.016a | |||||||
| Proportion of total effect that is mediated | 0.147 | 0.042 | 0.001 | −0.016 | 0.070 | 0.166 | |||||||
| Sobel test | −0.002a | −0.001a | −0.000 | −0.000 | −0.001a | −0.003a | |||||||
| KHB decomposition | |||||||||||||
|
|
Proportion of mediation effect (%) | 8.04 | 1.96 | —b | — | 5.74 | 10.98 | ||||||
aP<.001.
bAlcohol intake and physical exercise were not taken into KHB analysis as the mediation effects were not significant.