The aim of this study was to combine knowledge about how clinicians learn with a review of educational interventions to prevent delirium in hospitalised patients. The primary aim was to evaluate the effectiveness of approaches to delirium education. A detailed search of educational and medical databases was undertaken. The type of intervention used was classified according to the PRECEED model of Green and colleagues, using factors relevant to behaviour change in health promotion. The effectiveness of the intervention was determined by assessing changes in staff performance and patient outcomes. Nineteen studies of variable design and quality were identified. Studies using predisposing, enabling and reinforcing strategies together were more often effective in producing changes in staff behaviour and patient outcomes. Education and guidelines used together or in combination have little effect. When strategies to enable and reinforce changes in clinical practice are used together with education sessions, outcomes for patients are more positive.