Increasing evidence is pointing towards the efficacy of intervention programmes in decreasing the incidence of delirium among older people admitted to hospital. We have previously shown that an educational package directed at doctors and nurses significantly decreased the point prevalence of delirium among older people on a general medical ward. It is not yet established whether specific and fixed patients characteristics influence the rate of response to such an intervention. Methods A secondary, exploratory stratified analysis was carried out to determine whether age, sex and presence of dementia might influence the effect of the intervention through a delirium educational package. This information is important in order to increase the effectiveness of preventive measures across various patients subgroups. Results Male gender (OR 0.17, 0.05-0.65) significantly (p = 0.030) and positively influence the response to the educational package compared to female gender (OR 1.04, 0.38-2.81). Neither age nor the presence of underlying dementia was associated with a significant influence on the rate of response to the delirium prevention package. Conclusion Staff may be more likely to implement an earlier and more effective intervention for males who are perceived as presenting a higher risk to themselves and others. The results reported highlight the need to increase staff's vigilance to female patients whose prodromal symptoms of delirium may be underemphasised. Understanding different sub-group responses to preventive educational packages for delirium is an important consideration if these are to be applied widely. Copyright © 2006 John Wiley & Sons, Ltd.