Table 1.
Linear mixed models to estimate change in step count over seven days follow-up and the effect of treatment advice (SA = Stay active versus AA = Adjust activity) on this change
| N = 99 | Model 1: Change trajectory | Model 2: Effect of treatment advice | ||
|---|---|---|---|---|
| Fixed effects in model | Estimate (SE) | p-value | Estimate (SE) | p-value |
| Day 1 (intercept) | 4434 (405) | <0.001 | 4317 (558) | <0.001 |
| +Stay active advice | - | - | 243 (812) | 0.76 |
| Day (β, linear term) | 3160 (398) | <0.001 | 2773 (545) | <0.001 |
| +Stay active advice | - | - | 824 (796) | 0.30 |
| Day2 (β, quadratic term) | −987 (161) | <0.001 | −865 (220) | <0.001 |
| +Stay active advice | - | - | −257 (321) | 0.42 |
| Day3 (β, cubic term) | 94 (18) | <0.001 | 75 (24) | 0.002 |
| +Stay active advice | - | - | 34 (36) | 0.34 |
Fixed effects in models are presented with step count as outcome
Model 1 is the estimated change trajectory without the effect of treatment advice. Model 2 includes the effect of treatment advice, where the effect of the Adjust activity is presented first followed by the added effect of the Stay active advice
A third-order polynomial function was used with a linear term (Day) describing the initial increase (positive value), a quadratic term (Day2) describing the level-off of the initial increase (negative value) and a cubic term (Day3) for the additional final increase in step count over time