Implementation dosage and depressive symptoms following school-based mindfulness and relaxation: Complier Average Causal Effect estimation in the INSPIRE cluster randomised trial.

Implementation variability is hypothesized to be a key treatment effect modifier of universal, school-based mental health interventions, but the evidence base remains meagre. We used Complier Average Causal Effect (CACE) estimation to determine if dosage variability modified treatment effects of relaxation techniques and mindfulness-based exercises in the INSPIRE cluster randomised trial. Analyses were split by phase of education (mindfulness primary: k = 65 primary schools, N = 3,630 pupils; mindfulness secondary: k = 29 secondary schools, N = 3,708 pupils; relaxation primary: k = 66 primary schools, N = 3,800 pupils; relaxation secondary: k = 26 secondary schools, N = 3,252 pupils) and considered compliance effects at the 50th (moderate) and 75th (high) percentiles in the distribution of intervention dosage (total minutes delivered). Despite null intent to treat findings, we observed compliance effects for both interventions. In primary schools, moderate (320 min) and high (376 min) compliance to relaxation techniques significantly reduced depressive symptoms (d=-1.52 and d=-1.48, respectively). In secondary schools, moderate (207 min) and high (285 min) compliance to mindfulness-based exercises also significantly reduced depressive symptoms (d=-1.15 and d=-1.40, respectively). Contrastingly, in primary schools, moderate (325 min) and high (393 min) compliance to mindfulness-based exercises significantly increased depressive symptoms (d = 0.32 and d = 0.48, respectively). In secondary schools, moderate (180 min) and high (275 min) compliance to relaxation techniques also significantly increased depressive symptoms (d = 0.67 and d = 0.83, respectively). The effects of universal, school-based mental health interventions promoting relaxation and mindfulness are contingent on the delivery of a minimally effective dose. However, these interventions may not be appropriate in all educational contexts and could result in iatrogenic effects.Trial registration: Trial registration number: ISRCTN16386254. Trial registration date:04/09/2018.