Table 4 Illustration of the effect reduction to the mean when no adjustment for baseline outcome values is performed
Patient characterisation at baseline | Intervention’s effect (standard error) obtained with: | Consequence of the reduction to the mean | ||
|---|---|---|---|---|
N | No adjustment for baseline values | Baseline values as a covariate | ||
Middle condition | 141 | −0.107 (0.064) | −0.112 (0.067) | No major consequence |
Middle + worse condition | 173 | −0.036 (0.049) | −0.028 (0.054) | Effect overestimated |
Middle + better condition | 181 | −0.019 (0.066) | −0.047 (0.059) | Effect underestimated |