Universal cannabis outcomes from the Climate and Preventure (CAP) study: a cluster randomised controlled trial

Abstract

Background: The Climate and Preventure (CAP) study was the first trial to assess and demonstrate the effectiveness of a combined universal and selective approach for preventing alcohol use and related harms among adolescents. The current paper reports universal effects from the CAP study on cannabis-related outcomes over three years.

Methods: A cluster randomized controlled trial was conducted with 2190 students from twenty-six Australian high schools (mean age: 13.3 yrs., SD 0.48). Participants were randomised to one of four conditions; universal prevention for all students (Climate); selective prevention for high-risk students (Preventure); combined universal and selective prevention (Climate and Preventure; CAP); or health education as usual (Control). Participants were assessed at baseline, post intervention (6-9 months post baseline), and at 12-, 24- and 36-months, on measures of cannabis use, knowledge and related harms. This paper compares cannabis-related knowledge, harms and cannabis use in the Control, Climate and CAP groups as specified in the protocol, using multilevel mixed linear models to assess outcomes.

Results: Compared to Control, the Climate and CAP groups showed significantly greater increases in cannabis-related knowledge initially (p < 0.001), and had higher knowledge at the 6, 12 and 24-month follow-ups. There was no significant difference between the Climate and CAP groups. While no differences were detected between Control and the CAP and Climate groups on cannabis use or cannabis-related harms, the prevalence of these outcomes was lower than anticipated, possibly limiting power to detect intervention effects. Additional Bayesian analyses exploring confidence in accepting the null hypothesis showed there was insufficient evidence to conclude that the interventions had no effect, or to conclude that they had a meaningfully large effect.

Conclusions: Both the universal Climate and the combined CAP programs were effective in increasing cannabis-related knowledge for up to 2 years. The evidence was inconclusive regarding whether the interventions reduced cannabis use and cannabis-related harms. A longer-term follow-up will ascertain whether the interventions become effective in reducing these outcomes as adolescents transition into early adulthood.

Trial registration: This trial was registered with the Australian New Zealand Clinical Trials Registry (ACTRN12612000026820) on the 6th of January 2012, https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=347906&isReview=true.

Keywords: Cannabis; Combined; Online; Prevention; School; Selective; Universal.

Fig. 1

Power to detect differences in binary outcomes as a function of prevalence and effect size. Power was calculated using a method for three-level binary data randomized at the third level [34], using the function: $\mathrm{\phi }=\mathrm{\Phi }\left\{\frac{\left|p_1-p_0\right|\sqrt{N_1{N}_2{N}_3/f_3}-{\mathrm{\Phi }}^{-1}\left(1-\mathrm{\alpha }/2\right)\sqrt{2\overline{p}\left(1-\overline{p}\right)}}{\sqrt{p_0\left(1-p_0\right)+p_1\left(1-p_1\right)}}\right\}$ where N₁ is the size of the level 1 units (repeated observations of participants in the current study); N₂ the size of level 2 units (participants per school); N₃ the number of level 3 units (schools); p₁ and p₀ the prevalence of the outcome in the treatment and control groups respectively; $\overline{p}=\left(p_0+p_1\right)/2$; f₃ = 1 + N₁(N₂ – 1)ρ₂ + (N₁ – 1)ρ₁; ρ₁, ρ₂ the correlations within level 1 and level 2 units. Power calculations for the trial analyses were performed using N₁ = 5, N₂ = 80, N₃ = 5, ρ₁ = 0.36, ρ₂ = 0.03. Contour lines join regions with equal power in increments of 0.1, with the power of some contours labelled in white

Fig. 2

Predicted means of cannabis knowledge at each measurement occasion for each intervention group. Cannabis knowledge scores were on a scale from 0 to 16, with higher scores representing greater knowledge about cannabis. Black lines represent 95% confidence intervals for the predicted mean

Fig. 3

Predicted probabilities of cannabis use at each measurement occasion for each intervention group. A single survey item asked participants whether they had used cannabis in the past 6 months. Black lines represent 95% confidence intervals for the predicted probability

Fig. 4

Predicted probabilities of experiencing any harm from cannabis at each measurement occasion. Participants were asked whether they had experienced any of 6 different harms as a result of their cannabis use in the past 6 months. Black lines represent 95% confidence intervals for the predicted probability