Table 5 Aim-specific statistical analysis plan
Analysis | Individuals analyzed | Sample size | Estimated difference based on previous results mean (SD) | Power | Method |
|---|---|---|---|---|---|
Between group mean VSS-SF difference at w3, w6, m3, and m12 | All participants | 160+160 | 8.1 (7.4) vs 11.5 (9.9) [13] | 94% | ANCOVA |
Between group mean EQ-5D difference until m3 | All participants | 160+160 | Unknown | Unknown | Mixed models |
Between group mean EQ-5D difference at 12m | All participants | 160+160 | Unknown | Unknown | ANCOVA |
Between group mean DHI difference until m3 | All participants | 160+160 | 24.4 (20.8) vs 29.2 (21.1) [13] | 54% | Mixed models |
Mean of means T25-FW (seconds) and timed balance tests (seconds) until m3 | All participants | 160+160 | Unknown | Unknown | Mixed models |
Between group mean difference of means of vHIT gain (affected side) at w6 and m3 | Participants at sites with access to vHIT | 80+80 | 0.52 (0.24) vs 0.69 (0.25)a [31] | 99% | Ranked ANCOVA |
Proportion with catch-up saccades on vHIT (affected side) at w6 and m3 | Participants at sites with access to vHIT | 80+80 | Unknown | Unknown | Binary logistic regression |
Body sway until m3 | All participants | 160+160 | Unknown | Unknown | Mixed models |
Enhanced gait and body sway analysis until m3 | Umeå participants only | 25+25 | Unknown | Unknown | Mixed models |
Total steps since last visit at w6 and m3. | All participants | 160+160 | Unknown | Unknown | ANCOVA |
Self-reported vestibular rehabilitation compliance until 6w (mean number of training sessions) | All participants | 160+160 | Unknown | Unknown | Mixed models |
Proportion with falls or fractures | All participants | 160+160 | Unknown | Unknown | Binary logistic regression |
Proportion with BPPV | All participants | 320 | Estimated frequency 10% | N/A | Binary logistic regression |