Productive Failure and Learning about Epidemics and Complex Systems in Medical Education

Abstract

Over recent decades, global outbreaks such as AIDS, SARS, MERS, and COVID-19 have highlighted the importance of epidemiology for health and society. Despite advances in data-driven outbreak monitoring, epidemiology education often relies on teacher-centred methods that emphasise memorisation rather than explanatory understanding. Many epidemiological concepts are best understood as complex systems, where interactions between individuals produce emergent population-level behaviours. Traditional approaches leave many students struggling to link these dynamics with underlying mechanisms, creating a need for student-centred pedagogies.

This study examines Productive Failure (PF)—an approach where students first attempt problems before instruction—compared with Direct Instruction (DI), where instruction precedes problem solving. PF was combined with Agent-Based Models (ABMs), which allow learners to simulate disease spread and explore feedback mechanisms. To date, no studies have investigated the integration of PF and ABMs in epidemiology education.

A quasi-experimental design was implemented with 35 undergraduate medical students randomly assigned to PF (n=20) or DI (n=15) groups across three two-hour online sessions. Both groups engaged with identical challenge problems through ABMs, with instruction provided either after (PF) or before (DI). Pre- and post-tests assessed declarative and explanatory knowledge of complex systems, while challenge problems measured transfer to near and far domains. Focus-group interviews explored student experiences.

Findings showed no significant post-test differences between PF and DI for declarative or explanatory knowledge. However, PF students demonstrated stronger near-transfer gains from pre- to post-test and decisively outperformed DI students on far-transfer tasks, applying epidemiological knowledge to novel contexts.

The study suggests that while both approaches support knowledge development, PF combined with ABMs may be especially effective for enhancing students’ transfer of learning. Future research should replicate these findings with larger samples and explore applications in broader medical and public health contexts.