Table 3.
Multiple linear regression analyses.
| Variable | Model 1a | Model 2b | Model 3c | ||||
|
|
|
β (95% CI) | P value | β (95% CI) | P value | β (95% CI) | P value |
| Primary outcome analysis | |||||||
|
|
Digital literacy score | 1.99 (1.02 to 2.95) | <.001 | 1.91 (0.93 to 2.89) | <.001 | 1.90 (0.91 to 2.90) | <.001 |
| Domain-level analyses | |||||||
|
|
Social domain | 0.34 (0.09 to 0.61) | .009 | 0.35 (0.09 to 0.61) | .009 | 0.35 (0.08 to 0.61) | .01 |
|
|
Instrumental domain | 0.66 (0.23 to 1.10) | .003 | 0.71 (0.26 to 1.15) | .002 | 0.70 (0.24 to 1.15) | .003 |
|
|
Reassurance domain | 0.32 (0.06 to 0.59) | .02 | 0.30 (0.03 to 0.57) | .03 | 0.32 (0.05 to 0.60) | .02 |
|
|
Pastime domain | 0.30 (–0.02 to 0.62) | .06 | 0.29 (–0.05 to 0.64) | .10 | 0.28 (–0.06 to 0.63) | .11 |
aModel 1: group (control [reference group] / intervention), baseline domain score.
bModel 2: predictors in Model 1 and age, gender, education, housing type and living arrangement at baseline.
cModel 3: predictors in Model 2 and social isolation (Lubben Social Network Scale-6) and loneliness (University of California, Los Angeles 3-item loneliness scale) at baseline.