In collaboration with a bicycle airbag helmet company, data were collected to help explain events where head protections are deployed. The head protection records activations continuously, and when a head protection is deployed, this information is sent to the company. The company invited affected cyclists to (i) participate in a web survey, and (ii) share their data with researchers. The first aim of the study was to investigate the proportions between different severities of accidents, i.e., how many accidents with serious injuries occur for every accident with minor injuries, while the second aim was to predict when bicycle accidents will occur.

A total of 196 cyclists completed the web survey. In addition, head protection data were collected from 355 other cyclists, of which 264 had their helmet deployed. Web survey participants described the event and answered questions concerning the event. An event was defined as when the head protection was deployed. Not all events ended up in an accident. Helmet data contained information about the speed and location of the trips performed before and after an unique event, and the event trip as well.

One of the 182 cyclists ended up in hospital care. The accident data collected indicated the proportions of cyclists who needed hospital care (1= severe injuries), cyclists with injuries (15), slight injures (85) and cyclists who could continue as before (81 = no injuries). The head protection data confirmed the web survey findings, but also demonstrated that the head protection, on journeys that ended with head protection deployment, had a higher degree of activations (different head protection conditions) compared to journeys where it was not deployed. Furthermore, on trips made after deployment, the head protection had lower levels of activations, which can be understood as the cyclists adapting their behavior by, for example, riding more carefully (but not slower).