A group randomized control trial to test the efficacy of the Road to Mental Readiness (R2MR) program among Canadian military recruits

Power and sample size calculations were conducted with the Optimal Design Plus Empirical Evidence version 3.0 software [13] which is designed specifically for conducting power and sample size analyses in GRCTs. Required parameters including ICCs and average platoon (cluster) size were drawn initially from research leading up to the GRCT in our target population and updated [14] using data from the pilot phase of the GRCT. ICCs in the pilot GRCT ranged from 0 to 0.038. Average platoon size was estimated to be 28 taking into consideration a conservative participation rate of 50%. We assumed equal cluster size and 0 proportion of variance explained by group level covariates; Based on the existing literature, effect size was conservatively set at 0.2. The trial ended when estimated sample size was obtained. Further, this trial endpoint coincided with a significant change to BMQ training, a shortening of 2 weeks, which, had the trial continued to collect data, would have required a change in the third assessment point.

The randomization scheme took into consideration the fact that recruit platoons arrive at CFLRS in a staggered fashion over the course of the training year. For the 2016–17 and 2017–18 training years, the biostatistician was provided the training calendar several months ahead of time to generate the randomization scheme.

Intervention details have been taken from a previous publication on the R2MR program [18].

The trial procedures were piloted for feasibility with 10 platoons in late 2016, with close attention paid to randomization, scheduling, blinding, data collection (e.g., participation and retention rates), and data quality (e.g., missing data rates and patterns). Overall, the procedures were successfully implemented [23] minor modifications were made to data collection scripts to maximize participation and retention and to minimize missing data.

For measures with limited psychometric work or mixed results for construct/factorial validity (the CAF Mental Health Service Use Questionnaire (CAF-MHSUQ), the R2MR Program Evaluation Form for Mental Health Literacy (MHL), the Test of Performance Strategies (TOPS), and the Marlowe-Crown Social Desirability Scale (MC-SD)), we performed exploratory and confirmatory factor analyses with the pilot and the full trial data. Due to space limitations, these are captured in a separate report [24]. Scoring on these measures was based on the results from our psychometric analyses.