Table 4.
Fall prediction models with parameter coefficients and P values.
| Models | B | P value | Parameters |
| Model 1: questionnaires only | |||
|
|
-0.70 | <.001 | Constant |
|
|
0.46 | .004 | Fall history |
|
|
0.33 | .04 | GDS depression score |
| Model 2: add activities and gait characteristics’ medians | |||
|
|
-0.66 | <.001 | Constant |
|
|
0.67 | <.001 | Fall history |
|
|
0.41 | .04 | Low frequency percentage VT <0.7 Hz |
|
|
-0.43 | .02 | Lying duration |
|
|
0.86 | .002 | Sample entropy VT |
|
|
-0.53 | .03 | Sample entropy ML |
| Model 3: replace gait characteristics’ median if extreme has stronger association | |||
|
|
-0.68 | <.001 | Constant |
|
|
0.69 | <.001 | Fall history |
|
|
1.02 | <.001 | Sample entropy VT (Low extreme) |
|
|
0.61 | .002 | Low frequency percentage VT <0.7 Hz |
|
|
-0.60 | .004 | Lying duration |
|
|
-0.47 | .03 | Sample entropy ML |