Multinomial logit model of bicycle injury risk in Hong Kong

Background More and more countries have been promoting bicycle transport mode, as one of the key strategies in sustainable transport development. In many European and American countries (e.g. Netherlands, Denmark and Canada), bicycle has been a popular commuter transport mode. In contrast, bicycle is mainly used for recreational purpose in Hong Kong. Well planed bicycle transport infrastructure was absence, except smallscale and isolated bicycle path network in suburban area. Unfavorably, proportion of bicycle crash to total road crash has been increasing, from 5.3% in 1997 to 12.7% in 2004. Therefore, factors contributing to bicycle injury risk are worth exploring.

Study Design An integrated database, Road Casualty Information System (RoCIS), was jointly developed by the Police, Transport Department and hospital, linking up the trauma records of casualties admitted to the accident and emergency department of hospital and the crash records maintained by the Police and Tranport Department the advantage of RoCIS, information on bicyclist demographics, injury severity, injured body part, collision type, road design and vehicle attributes, of 682 bicycle casualties during the period 20042006 have been collected.

Results In this study, multinomial logit regression is applied to identify the signifi cant factors contributing to the risk of severe and lifethreatening injury of bicycle casualties.

Results indicate that middle age and elderly bicycle casualties are more likely to have severe injury. Besides, bicycle casualties with severe head injury and with motor vehicles involved are remarkably more likely to have lifethreatening injuries. Unfortunately, the helmet wearing rate is extrmely low and at 2% only. Therefore, safety education, campaign and enforcement could be targeted to the middle age and elderly bicyclists. In particular, use of protective device and compliance to traffi c rules should be promoted. Also, access of bicycle on the motorway should be scrutinized.

Acknowledgement We gratefully acknowledge Tuen Mun Hospital for providing the casualty data, and the Road Safety and Standards Division of the Transport Department of the Hong Kong Special Administrative Region (HKSAR) for providing the traffi c accident data, for the research work that is reported in this paper. The views expressed are those of the authors and do not represent January 2013 ISSN 2157-118X 77 those of the HKSAR goverment. The work that is described in this paper was supported by a grant from the Research Grants Council of the Hong Kong Special Administrative Region, China (Project No.: HKU7176/07E), the Engineering Postdoctoral Fellow Programme, and an Outstanding Rsearcher Award from the University of Hong Kong.