Bicycle Travel, Injury Risk and Conspicuity. Application of Epidemiological Methods

Abstract

Background Regular cycling provides health, environmental and economic benefits but is rarely part of everyday travel for many people. The real and perceived risk of injury is commonly cited as a significant deterrent to engaging in such activity and is sometimes associated with the danger of motorised traffic. Bicycle-motor vehicle collisions often result from the driver’s failure to detect the cyclist in time, suggesting that conspicuity may be a contributing factor. Conspicuity can be categorised into physical conspicuity and attention conspicuity. The former may be enhanced by using conspicuity aids whereas the latter may be improved through the safety in numbers effect. In New Zealand, one of the most car-dependent countries, bicycling is marginalised and receives relatively little attention in the transport and road safety research agenda. As a result, there is insufficient evidence regarding bicycle travel, injury risk and associated factors to guide education and policy. Aims This thesis comprises three main components – bicycle travel, injury risk and conspicuity and aims:  to investigate the prevalence of bicycle travel, particularly for a trip to work, and identify associated enablers and barriers,  to estimate the exposure-based rate of bicycle crash injuries and identify associated risk and protective factors, and  to assess the role of conspicuity in bicycle crashes involving a motor vehicle. Methods This research was based on the Taupo Bicycle Study, a prospective cohort study involving 2628 adult cyclists (aged 16 years and over). The participants were recruited from the Lake Taupo Cycle Challenge (New Zealand’s largest mass cycling event) in 2006 (N=2438) and in 2008 (N=190) using a web-based questionnaire. The questionnaire asked about demographic characteristics, general cycling activity and crash experience in the past twelve months and habitual use of injury preventive measures with options ranging from never to always. Bicycle crash outcome data were collected through linkage to insurance claims, hospital discharges, mortality records and police reports and covered the period between recruitment and 30 June 2011 (i.e., a median follow-up of 4.6 years). As only a small number of bicycle-motor vehicle crashes were identified in the linked data, another linkage was undertaken in 2013 and the longer-term outcome data covering the period till 30 April 2013 (a median follow up of 6.4 years) were used in the analysis addressing the third research question. For all longitudinal analyses, the study sample was restricted to 2590 cyclists who were resident in New Zealand at recruitment. Cox proportional hazards regression modelling for repeated events was performed using a counting process approach and factors influencing the likelihood of experiencing crash episodes were identified. The investigations of the Taupo Bicycle Study were supplemented with analyses of national data (aggregate or de-identified) obtained from Censuses, Household Travel Surveys, hospital discharge and mortality records, police reports, cycle counts and climate database. Results Bicycle travel: Analysis of Census data from 1986 to 2006 showed a decline in the prevalence of cycling to work from 5.6% to 2.5%, respectively. About one-third of the Taupo Bicycle Study participants reported cycling to work at least once a week. The prevalence was lower in women, older age groups, non-Māori, less experienced cyclists and in the Auckland region. The participants of the Taupo Bicycle Study reported (in decreasing order of frequency) that more bicycle lanes, more bicycle paths, better bicycle security in public places, reduced vehicle speed and bike friendly public transport as important factors that would encourage them to cycle more often. They reported (in decreasing order of frequency) that access to shower facilities at work, the need to negotiate fewer difficult intersections, rising costs of petrol, fewer car parks, bikes designed to commute, and rising costs of car parking would be important incentives to cycling to work. They perceived that adverse weather and the danger of car-dominated transport environment including road safety, traffic en route, driver attitude and behaviour and breathing polluted air were the most important barriers to bicycle travel. Analysis of automatic cycle counts on Tamaki Drive in Auckland and weather data showed that selected weather variables such as gust speed, rain, temperature and sunshine duration accounted for 23% of the variance in hourly cycle volume and 56% of the variance in daily cycle volume. Injury risk: Analysis of hospital discharge, mortality and travel survey data revealed an increase in the number of on-road cycling injuries per million hours spent cycling from 21 injuries in 1996/99 to 31 injuries in 2003/07. The rate was higher than other types of road users except motorcyclists. During a median follow-up of 4.6 years, the participants of the Taupo Bicycle Study experienced 116 crashes that came to the attention of medical personnel or police per 1000 person-years, of which 66 occurred on the road and 10 involved a collision with a motor vehicle. This corresponds to 391 crashes per million hours spent cycling in general, and 240 crashes and 38 collisions per million hours spent cycling on the road. The crash risk was associated with age, body mass index, urban residence, region, cycling off-road, in the dark or in a bunch, type of bicycle used and prior crash history, with variations in effect estimates by crash type (on-road, off-road and collision crashes). In the Taupo Bicycle Study, participants who reported a history of bicycle crashes in the preceding year had a higher risk of involvement in future bicycle crashes (hazard ratio: 1.43; 95% CI: 1.28, 1.60). When this association was investigated using a propensity score stratification approach, it was significant only in the highest two quintiles of the propensity score where the likelihood of having experienced a crash was more than 33%. The association was also stronger for previous crashes that had received medical care compared to those that had not. In the Taupo Bicycle Study, participants usually residing in Auckland had a higher risk of on-road bicycle crashes (hazard ratio: 1.47; 95% CI: 1.22, 1.76) but had a similar risk of off-road bicycle crashes (hazard ratio: 0.96 (95% CI: 0.77, 1.21) compared to the participants from the other regions of New Zealand. In the mediation analysis, about 53% of the increased risk in Auckland was explained by differences in the participants’ cycling patterns and urban residence. It is likely that the remaining risk differential is at least partly influenced by factors associated with Auckland’s crowded, car-dominated transport environment. Conspicuity: Analysis of hospital discharge, mortality and travel survey data to explore the effect of attention conspicuity showed regional differences in the rate of on-road cycling injuries in relation to the level of car and bicycle use, supporting the safety in numbers (or risk in scarcity) effect. Analysis of the Taupo Bicycle Study data to investigate the effect of physical conspicuity found that the risk of bicycle crashes involving a motor vehicle was similar across different usage patterns of conspicuity aids. In subgroup analyses, the most conspicuous group had a higher risk in Auckland but a lower risk in other regions. Conclusions In New Zealand, the prevalence of bicycle commuting is low and has been in decline between 1986 and 2006 (note that there are signs of recovery recently). The exposure-based rate of bicycle crash injuries is relatively high compared to other road user categories. Regional differences in travel patterns and injury risks suggest that the risk in scarcity effect exists for New Zealand cyclists. This may be attributed partly to the poor attention conspicuity of cyclists and partly to poor cycling infrastructure in the country’s car-dominated traffic environment. Conspicuity aids used to enhance physical conspicuity may be ineffective in such situation. In the last two decades, New Zealand has been caught in a vicious circle – a lower proportion of cyclists on the road decreases their conspicuity and poses them a higher crash risk which in turn discourages bicycle use. Turning this vicious circle to a virtuous one requires cooperative and multidisplinary efforts to promote cyclists’ safety and encourage cycling on New Zealand roads.