Abstract Background Final year health sciences’ students work in clinical teams. Medical errors in are associated with teamwork and communication failures. Non-technical skills performance gaps are related to medical error. Little is known about the metacognitive strategies of final year clinical students and about how their mental models develop and evolve. Research in non-technical skills has focused on performance, and not on the underlying causes of actions such as mental models. Methods We conducted a mixed, parallel convergent study with final year medicine and nursing students. In the instructional design, leadership skills (maintain active listening, give clear commands, allow others to participate in decision-making, prioritize essential tasks) and metacognitive strategies (planning, monitoring and control) were included in simulated medical emergency scenarios with participants performance measurements for four weeks. Results The median of the initial performance was 25 (IQR: 25), in the second week it was 25 (IQR: 6–50), in the third week it was 50 (IQR: 31–50), and in the fourth week it was 75 (IQR: 75–100). No significant difference was found in performance by program, gender or age. Performance significantly improved by the fourth week of training (p: <0.0001). Conceptual mastery was achieved after the first week, multidimensional mental models went from being inhibiting to being enabling between the third and fourth week. Conclusions The conceptual domain of leadership is not associated with performance improvement; individual and team performance improves when multidimensional mental models evolve from being inhibiting and become enabling, possibly catalyzed by the conscious reflection and the mastery of metacognitive strategies.