Table 4.
Associations between website usage variables (binary) and weight changea in the final cohort (n=1258).
| N (%) | Coefficient | 95% CI | P value | |
| At least one food entry day | 893 (71.0) | -0.35 kg | -1.28 kg to 0.58 kg | .47 |
| At least one exercise entry day | 677 (53.8) | -0.67 kg | -1.52 kg to 0.18 kg | .12 |
| At least one exercise minute | 741 (58.9) | -0.61 kg | -1.47 kg to 0.25 kg | .16 |
| At least one forum post | 291 (23.1) | -1.55 kg | -2.55 kg to -0.55 kg | .002 |
| At least one Friend | 547 (43.5) | -0.80 kg | -1.66 kg to 0.06 kg | .07 |
a Website usage variables were coded as binary (at least once versus never). Analyses were adjusted for baseline BMI. Because all members in the sample had two or more log-ins and weigh-ins, and most members had at least one SparkPoint, those variables were not included.