71 Mindfulness-Based Stress Reduction for elementary school teachers: a randomized controlled trial Secondary outcomes Well-being Analyses of secondary outcomes showed significant group by time interactions for almost all secondary outcomes, with all effects being in favor of the intervention group. With regard to the mental health outcome measure of well-being, moderate between-group effect sizes at post-treatment and follow-up were found for well-being. Proximal secondary outcomes Analyses of proximal secondary outcomes showed significant group by time interactions for all proximal outcomes; mindfulness skills, self-compassion and emotion regulation. All effects were in favor of the intervention group. Large effects were found for mindfulness skills and self-compassion at post-treatment and follow-up, and for emotion regulation at follow-up. Moderate between-group effect sizes were found for emotion regulation at post-treatment. Between-group post-hoc tests were significant for all proximal outcomes at all timepoints (p’s < 0.05). Per-protocol analyses using completers only did not reveal different results compared to ITT analyses for the proximal outcomes. Distal secondary outcomes Regarding distal secondary outcome measures, large effects were found for teacher self-efficacy at post-treatment and follow-up and moderate effect sizes at both time points for teacher-perceived classroom climate quality. Between-group post-hoc tests were significant for almost all distal outcomes at all timepoints (p’s < 0.05), except for classroom climate quality at post-treatment, which just fell short of significance (p = 0.08). No significant interaction effect was found for teacher-perceived pupil-teacher relationship (p = 0.15), and the between-group post-hoc tests did indicate that there was no significant between-group difference at any of the timepoint (p’s > 0.10). Per-protocol analyses using completers only did not reveal different results compared to ITT analyses for the distal outcomes. Sensitivity analyses Adjusting the models for covariates did not change conclusions drawn from the models for primary and secondary outcomes. None of the covariates was significantly related with the outcome variable, and a significant group by time interactions were still found for all outcomes (p’s < 0.01), except for the effect on the pupil-teacher relationship (p = 0.15). One outlier was identified for the primary outcome perceived stress at post-treatment. This person scored at least three standard deviations higher than the mean. However, excluding the data of this person from the analyses did not substantially change the results of the linear mixed model and between-group post-hoc tests. Also, one outlier was identified for teacher’s self-efficacy at follow-up, class climate at post-measurement and pupil-teacher relationship at post-measurement and follow-up. In all cases, the scores 4
RkJQdWJsaXNoZXIy MTk4NDMw