Background. Research indicates that school-based first aid programs appear to improve students' knowledge and skills. However, evidence for their effectiveness is limited by a lack of rigorously designed studies. This research used a cluster randomised crossover trial to assess the effects of two different types of first aid training on the frequency and appropriateness of older adolescents' first aid behaviours towards their peers 12 months after training. Methods. Schools eligible to participate were government funded and able to accommodate first aid training and survey time for two consecutive Year 10 student cohorts. Four Australian public schools were matched in two pairs and randomly assigned to receive either physical first aid (PFA) or teen mental health first aid (tMHFA) training for their Year 10 student cohort (mean age 16 years). In the second year, the new Year 10 cohort received the other intervention. Four cohorts were randomised to receive PFA and four were randomised to receive tMHFA. Online surveys were administered at baseline and 12 months after training, measuring whether students had encountered a peer needing PFA, whether they had provided PFA, what actions they performed and, if applicable, why they had been unable to help the person. Only research staff analysing the data could be blinded to measurement occasion, school identity and condition. Results. Four cohorts received PFA and four received tMHFA. The results indicated that there were no differences between groups regarding the frequency of appropriate first aid actions performed 12 months after training. The most common types of PFA provided to a peer were sending for help and wound care. Students most commonly mentioned someone else attending to their peer or lacking skills or experience as reasons for not performing PFA actions. Conclusions. More research that examines first aid behaviours using rigorous, longitudinal study designs is needed to establish the effectiveness of school-based first aid training for older adolescents.