In its latest global assessment of road safety, the World Health Organization (WHO) reported that half of the 1.2 million fatalities occurring each year on the world's roads concern vulnerable road users (VRUs), with children and the elderly being overrepresented among victims 
. Pedestrians, pedal cyclists, and motor cyclists are considered as vulnerable since they benefit from little or no external protective devices that would absorb energy in a collision. They constitute with almost no exception the weak party in a road traffic crash.
Beyond preventive policies such as traffic-calming measures, enforcement of drunk driving legislation and increased visibility/conspicuity 
, individual passive safety is also essential to protect VRUs from injury. As far as motorised two-wheelers are concerned, the most effective protection that can be offered is the helmet. Evidence from a systematic review shows that it reduces the risk of fatal injuries by 42% 
. When it comes to cyclists, authors of a Cochrane Collaboration analysed five case-control studies and concluded that helmets reduce by 63% to 88% the risk of head, brain, and severe injuries among cyclists. They concluded that bicycle riders of all ages should be encouraged to wear helmets 
. According to the European Transport Safety Council 
, the death risk per 100 million person kilometres travelled is 5.4 for bicyclists in Europe, compared with 0.7 for car users and 0.07 for bus and coach passengers. Because about half of fatal and serious injuries among bicyclists are head injuries, helmet use is recommended 
and sometimes even compulsory in a small number of countries.
There is, however, still no consensus on the best policy regarding helmets for bicyclists. First, the accuracy of evaluating studies is debated. When sources of bias are controlled for, the protective effects attributed to bicycle helmets 
become smaller than originally estimated 
. Second, it has been argued that bicyclists' helmet use has the potential to make other drivers less cautious concerning cyclists 
and to increase cyclists' propensity to take more risks (a phenomenon called risk compensation or risk homeostasis) 
. Both hypotheses, however, rely upon little observational evidence. Finally, there is also a debate on whether helmet use should be compulsory 
. Some authors state that a helmet law might deter people from cycling, and thus diminish the benefits to health of regular exercise provided by cycling, 
. Such a decrease in cycling may also increase cyclists' vulnerability due to the lower awareness of this population by other road users 
Effective strategies to promote bicyclist helmet use are therefore still to be identified. These may include health education programmes, subsidised or free helmet distribution programmes, media campaigns, or interventions that include elements of the above. Community-based interventions and those providing free helmets have an effect on reported use, but were mostly aimed at children 
Identifying successful interventions should also provide an opportunity to understand the determinants of such protective behaviour. Psychosocial models such as the Theory of Planned Behaviour (TPB) 
, the Locus of Control, and the Health Belief Model 
provide a theoretical framework that has frequently been applied to study the determinants of behavioural intention. Among these models, TPB is of particular interest since all its components were shown to correlate with the intention to use a helmet among adolescents 
. These are attitudes toward the behaviour (affective and instrumental evaluations of performing the behaviour), subjective norms (perceived social pressure to perform a behaviour or not), and perceived behavioural control (the perceived ease or difficulty of performing the behaviour). Available studies suggest that the main determinant of helmet use is cyclists' motivation 
. All such studies are, however, hampered by substantial uncertainty regarding the accuracy of self-reported behaviour, which is potentially subject to bias such as social desirability 
and recall bias. 
In the present study, we aimed to assess the influence of two strategies for promotion of helmet use, using a randomised controlled protocol, and with direct observation of behaviour by a semi-automatic video system. Factors influencing adoption of helmet use were elicited with analyses based on data from a questionnaire completed by participants at inclusion.