Table 4.
Associations of various website features by tertile of use with different weight outcomes at end of data collection
| Top Tertile | Middle Tertile | Bottom Tertile | P valuea | ||
| Phase II weight change (regain) in kg, mean (SD) | |||||
| Log-ins | 4.0 (10.1) | 5.5 (9.4) | 6.9 (9.6) | P = .001 | |
| Minutes | 3.1 (6.1) | 6.2 (9.6) | 6.9 (8.6) | P < .001 | |
| Weight entries | 4.8 (6.1) | 4.3 (9.2) | 7.2 (9.3) | P = .002 | |
| Exercise entries | 3.6 (10.5) | 5.7 (9.3) | 7.0 (8.8) | P < .001 | |
| Added use after weight entry | 4.3 (10.3) | 4.9 (9.3) | 7.1 (9.6) | P = .002 | |
| Bulletin board posts | 4.8 (13.2) | 6.5 (10.2) | 5.3 (7.1) | P = .10 | |
| Bulletin board reads | 4.7 (12.1) | 6.2 (8.3) | 5.4 (8.2) | P = .16 | |
| Proportion of initial weight loss regained, mean (SD) | |||||
| Log-ins | 48.8% (127.5%) | 68.2% (98.9%) | 96.4% (134.7%) | P < .001 | |
| Minutes | 36.3% (58.2%) | 80.7% (137.5%) | 95.7% (120.1%) | P < .001 | |
| Weight entries | 55.4% (73.0%) | 56.4% (107.1%) | 101.0% (121.9%) | P < .001 | |
| Exercise entries | 43.0% (108.6%) | 78.5% (138.3%) | 91.6% (113.5%) | P < .001 | |
| Added use after weight entry | 54.1% (132.0%) | 59.8% (96.1%) | 99.3% (137.5%) | P < .001 | |
| Bulletin board posts | 60.0% (159.0%) | 88.4% (146.3%) | 69.9% (90.9%) | P = .02 | |
| Bulletin board reads | 55.6% (136.0%) | 82.1% (106.6%) | 75.3% (115.2%) | P = .01 | |
| Proportion of participants at least 4 kg below initial weight | |||||
| Log-ins | 58.3% | 39.5% | 24.5% | P < .001 | |
| Minutes | 59.7% | 36.9% | 26.0% | P < .001 | |
| Weight entries | 49.1% | 48.9% | 24.6% | P = .005 | |
| Exercise entries | 61.2% | 35.0% | 26.2% | P < .001 | |
| Added use after weight entry | 54.8% | 42.4% | 25.2% | P = .002 | |
| Bulletin board posts | 47.4% | 30.3% | 41.6% | P = .05 | |
| Bulletin board reads | 53.9% | 29.2% | 39.5% | P < .001 | |
a Two-tail P values based on logistic (binary) or linear (continuous) regressions analysis adjusted for race, sex, education, income, and age