UBC Faculty Research and Publications

Personal and trip characteristics associated with safety equipment use by injured adult bicyclists: a… Teschke, Kay; Brubacher, Jeff R; Friedman, Steven M; Cripton, Peter A; Harris, M A; Reynolds, Conor C; Shen, Hui; Monro, Melody; Hunte, Garth; Chipman, Mary; Cusimano, Michael D; Lea, Nancy S; Babul, Shelina; Winters, Meghan Sep 11, 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12889_2012_Article_4449.pdf [ 201.75kB ]
JSON: 52383-1.0221387.json
JSON-LD: 52383-1.0221387-ld.json
RDF/XML (Pretty): 52383-1.0221387-rdf.xml
RDF/JSON: 52383-1.0221387-rdf.json
Turtle: 52383-1.0221387-turtle.txt
N-Triples: 52383-1.0221387-rdf-ntriples.txt
Original Record: 52383-1.0221387-source.json
Full Text

Full Text

RESEARCH ARTICLE Open AccessPersonal and trip characteristics associated withsafety equipment use by injured adult bicyclists: across-sectional studyKay Teschke1*, Jeff R Brubacher2, Steven M Friedman3, Peter A Cripton4, M Anne Harris5, Conor CO Reynolds6,Hui Shen1, Melody Monro1, Garth Hunte2, Mary Chipman7, Michael D Cusimano7, Nancy Smith Lea8,Shelina Babul9 and Meghan Winters10AbstractBackground: The aim of this study was to estimate use of helmets, lights, and visible clothing among cyclists andto examine trip and personal characteristics associated with their use.Methods: Using data from a study of transportation infrastructure and injuries to 690 adult cyclists in Toronto andVancouver, Canada, we examined the proportion who used bike lights, conspicuous clothing on the torso, andhelmets on their injury trip. Multiple logistic regression was used to examine associations between personal andtrip characteristics and each type of safety equipment.Results: Bike lights were the least frequently used (20% of all trips) although they were used on 77% of trips atnight. Conspicuous clothing (white, yellow, orange, red) was worn on 33% of trips. Helmets were used on 69% oftrips, 76% in Vancouver where adult helmet use is required by law and 59% in Toronto where it is not. Factorspositively associated with bike light use included night, dawn and dusk trips, poor weather conditions, weekdaytrips, male sex, and helmet use. Factors positively associated with conspicuous clothing use included good weatherconditions, older age, and more frequent cycling. Factors positively associated with helmet use included bike lightuse, longer trip distances, hybrid bike type, not using alcohol in the 6 hours prior to the trip, female sex, older age,higher income, and higher education.Conclusions: In two of Canada’s largest cities, helmets were the most widely used safety equipment. Measures toincrease use of visibility aids on both daytime and night-time cycling trips may help prevent crashes.Keywords: Active transport, Bicycle safety, Visibility, Bicycle helmetBackgroundBicycling injuries are a concern both because of the dir-ect harm they cause individuals and because concernsabout safety are a deterrent to use of this healthy modeof transportation, especially in North America [1-3]. Inthe United States, collisions with motor vehicles resultin about 700 fatalities and 48,000 police-reported injur-ies per year among cyclists [4]. In Canada, with a popu-lation about one-tenth of the US, collisions with motorvehicles result in about 50 fatalities and 450 seriousinjuries (requiring hospitalization) per year amongcyclists [5]. These data do not account for all injuries tocyclists, since they do not tally crashes not involvingmotor vehicles, and they may also miss some that do [6].As outlined in William Haddon’s original work on traf-fic injury epidemiology, there are many potentialapproaches to injury reduction. Measures can be direc-ted at the individual cyclist or the cycling environment,and can be focussed on pre-event prevention or post-event mitigation [7]. A number of authors have noted adifference in emphasis in bicycling safety between north-ern Europe and North America, with an environmentfocus dominant in the former and an individual focusdominant in the latter [8,9]. Within individual-based* Correspondence: kay.teschke@ubc.ca1School of Population and Public Health, 2206 East Mall, University of BritishColumbia, Vancouver, BC, CanadaFull list of author information is available at the end of the article© 2012 Teschke et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the CreativeCommons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly cited.Teschke et al. BMC Public Health 2012, 12:765http://www.biomedcentral.com/1471-2458/12/765safety measures, there is a range of possible options, in-cluding those aimed at crash prevention (e.g., lights) andthose focussed on injury mitigation (e.g., helmets). Anumber of studies have measured use of individual-based safety equipment by cyclists [10-25], though fewhave documented use of multiple types of equipment inthe same population. Fewer still have examined charac-teristics (e.g., weather conditions, cyclist age) associatedwith use of safety equipment [10-12,14-18,20,22,25]. Bet-ter understanding could help inform priorities for im-provement, interventions to improve uptake, and therelative safety potential of these individual-based mea-sures vis-à-vis population-based alternatives such asbicycle-dedicated infrastructure.As part of a study of 690 cyclists injured in two ofCanada’s largest cities, Toronto and Vancouver, we col-lected data on use of lights, high conspicuity clothing onthe torso, and helmets. In addition, we collected data ontrip and personal characteristics that allowed us toexamine factors associated with use of these types ofsafety equipment.MethodsThe study methods were reviewed and approved by thehuman subjects ethics review boards of the University ofBritish Columbia, the University of Toronto, St. Paul’sHospital, Vancouver General Hospital, St. Michael’s Hos-pital, and the University Health Network (Toronto Gen-eral Hospital and Toronto Western Hospital). Allparticipants gave informed consent before taking part inthe study.Methods of study conduct have been described in de-tail elsewhere [26]. The study population consisted ofadult (≥ 19 years) residents of Toronto and Vancouverwho were injured while riding a bicycle in the city andtreated within 24 hours in the emergency departmentsof the hospitals listed above between May 18, 2008 andNovember 30, 2009.Eligible participants were interviewed in person bytrained interviewers, using a structured questionnaire(http://cyclingincities.spph.ubc.ca/files/2011/10/Inter-viewFormFinal.pdf ) as soon as possible after the injuryto maximize recall. Questions related to safety equip-ment use were the following (asked in this order): Did you have a back light that was turned on duringthis trip? Did you have a front light that was turned on duringthis trip? What colour was the clothing on your upper body? What colour was the helmet you were wearing?Questions about front and back lights were combinedand if at least one light was turned on, assigned a “yes”.The following torso clothing colours were classified ashighly visible based on evidence of conspicuity from thestudy of Hagel et al. [18]: white, yellow, orange, and red.Those who reported a helmet colour were classified aswearing a helmet. “Don’t know” or “refused” responsesfor all questions were grouped with the “no” category, toprovide a conservative estimate of the prevalences ofsafety equipment use. Only highly visible clothing on thetorso had large numbers of “don’t know” responses; werepeated analyses (removing “don’t know” responses) todetermine whether this conservative classification hadan impact on the results.The interview also collected data on trip characteris-tics (weather conditions, time of day, day of week, sea-son, trip distance, trip purpose, bike type used, alcoholuse in the 6 hours prior to the trip, drug use in the6 hours prior, sleep in the 24 hours prior, and whetherthe participant was cycling with a companion) and per-sonal characteristics (age, sex, education, income, em-ployment, cycling frequency, and whether theparticipant had a driver’s license, was an experiencedcyclist, had taken a cycling training course, and had chil-dren in the household).Unconditional multiple logistic regression was used toexamine associations between the use of each type ofsafety equipment (bike lights, highly visible clothing onthe torso, or helmet; as dependent variables) and the fol-lowing independent variables: city; all trip and personalcharacteristics listed above; and the two types of safetyequipment that were not the dependent variable in theanalysis. We used backwards selection to construct mul-tiple logistic regression models, starting by offering allvariables of interest. Based on results of the Wald testfor each variable, the variable with the highest p-valuewas removed and the model refit with the remainingvariables until all variables in the model were statisticallysignificant at the p < 0.05 level. Data analyses were per-formed using SAS 9.2 (SAS Institute, Cary, NC). In thispaper we present the unadjusted and adjusted results forthe variables in the final models. The results for fullmodels with all variables included (prior to backwardsselection) are available from the authors.ResultsDetails on the recruitment process are available else-where [26]. In brief, 2,335 injured cyclists attended oneof the five study emergency departments during thestudy period. Of these, 927 were deemed ineligible, 741deemed eligible and 690 participated (414 in Vancouver,276 in Toronto). There were 667 with unknown eligibil-ity (543 not contacted, 124 refusals). Participants repre-sented 93.1% of those confirmed to be eligible and 66.5%of those estimated to be eligible (based on the propor-tion eligible among those contacted). The most commonTeschke et al. BMC Public Health 2012, 12:765 Page 2 of 9http://www.biomedcentral.com/1471-2458/12/765reasons for ineligibility were not being a resident of thestudy city and being injured outside the city.Table 1 lists selected participant and trip characteris-tics. Most participants were men, younger than 40, welleducated, employed, earned more than $50,000 a year,were regular cyclists, and had a driver’s license. Most ofthe injury trips were short and utilitarian in nature. Fewparticipants had taken alcohol or drugs in the 6 hoursprior to the trip, were sleep deprived, or were travellingwith a companion. Most of the injury events were colli-sions (i.e., involved hitting a vehicle, object, surface,person or animal) rather than falls, and almost halfinvolved a motor vehicle (one-third directly and 14%indirectly in avoidance manoeuvres). Most occurred atnon-intersection locations.Use of safety equipment is outlined in Table 2. Light-ing was the least frequently used, with 135 (19.6%) parti-cipants indicating they had at least one light turned on,including 96 using both lights, 25 with only the backlight on and 14 with only the front light on. Seven parti-cipants responded “don’t know”. Responses about cloth-ing indicated that 230 (33.3%) wore white, yellow,orange or red on their torso. There were 56 participantswho responded “don’t know” and three who were notwearing clothing on the torso. We did not directly askfor information about use of reflective material, which isvisible at night under illumination. It was self-reportedby 59 individuals, 34 of whom were wearing colours thatwere not classified as conspicuous. Because this informa-tion was not directly solicited in questioning, reflectivematerial use is likely to have been under-reported. Hel-mets were the most frequently used safety equipmentwith 478 (69.3%) participants indicating their helmetcolour. Two responded “don’t know” and one personrefused to answer.Table 3 shows the logistic regression results for factorsassociated with having at least one bike light turned on.The strongest relationships were for time of day: only6% of participants had lights on during the daytime, ver-sus 44% at dawn or dusk, and 77% at night. Dull weatherTable 1 Characteristics of the study participants and thebicycling trips when they were injured (N=690)Characteristic Number (%)Male 410 (59.4%)Female 280 (40.6%)Age (of N = 685 reporting)19 to 29 years 250 (36.5%)30 to 39 years 177 (25.8%)40 to 49 years 108 (15.8%)50 to 59 years 91 (13.3%)60 to 69 years 49 (7.2%)≥ 70 years 10 (1.5%)Completed post-secondary diploma or degree 518 (75.1%)Employed 546 (79.1%)Income greater than $50,000 (of N = 610 reporting) 341 (55.9%)Had children in their household 104 (15.1%)Regular cyclist (cycled≥ 52 times per year) 608 (88.1%)Considered themselves an experienced cyclist 529 (76.7%)Had taken an urban cycling training course 42 (6.1%)Had bike maintained in the last 6 months 525 (76.1%)Had a driver’s license 620 (89.9%)Trip < 5 km 470 (68.1%)Trip purposeCommute to or from work or school 287 (41.6%)For exercise or recreation 177 (25.7%)For social reasons (e.g., movies, visit friends) 159 (23.0%)For personal business (e.g., shopping, doctor’s visit) 126 (18.3%)During work 17 (2.5%)Alcohol or drug use in 6 hours prior to tripAlcohol 73 (10.5%)Medications 52 (7.5%)Recreational drugs 25 (3.6%)Had less than 6 hours of sleep in 24 hours prior to trip 23 (3.3%)Cycling with a companion 109 (15.8%)Injury circumstancesCollision 497 (72.0%)Fall 193 (28.0%)Motor vehicle involved 331 (48.0%)Crash at an intersection 211 (30.6%)Table 2 Safety equipment use on the trip: bike lights;visible clothing; and helmetsAt least one bikelight turned onHighly visible clothingworn on the torsoHelmetwornYes/No* Yes/No Yes/NoAt least one bike light turned onYes 135/0 49/86 102/33No 0/555 181/374 376/179Highly visible clothing worn on the torsoYes 49/181 230/0 169/61No 86/374 0/460 309/151Helmet wornYes 102/376 169/309 478/0No 33/179 61/151 0/212* Note that all “no” categories also include those who didn’t know or refusedto answer the question;• for “at least one bike light turned on”, 7 participants (1.01%) indicatedthey didn’t know;• for “highly visible clothing worn on the torso”, 56 indicated they didn’tknow (8.1%), and;• for “helmet worn”, 2 indicated they didn’t know and 1 refused (0.44%);Teschke et al. BMC Public Health 2012, 12:765 Page 3 of 9http://www.biomedcentral.com/1471-2458/12/765also prompted light use, with strong associations forcloud cover, fog, mist, rain or snow. Those on weekendtrips and women were less likely to use lights (even afteradjusting for weather and time of day). Helmet use waspositively associated with use of lights.Few variables showed associations with wearing highlyvisible clothing on the torso (Table 4). Cold, wet weatherwas associated with lower odds of wearing conspicuousclothing. Older adults (50 to 59 years) and those whowere more frequent cyclists were more likely to wearsuch clothing. In a repeat of the analyses for highly vis-ible clothing, excluding the 56 participants who couldnot recall the colour of the clothing they wore, the vari-ables associated and their odds ratios and confidenceintervals were nearly identical to the adjusted analysesreported in Table 4.An array of variables were associated with helmet use(Table 5). Participants in Toronto, where there is nolegal requirement for adults to wear helmets, were lesslikely to wear them (59%) than those in Vancouver(76%). Trip characteristics positively associated with hel-met use included use of a bike light, longer trip dis-tances, and use of a hybrid style of bicycle. Cruiser bikeuse and consumption of alcohol in the 6 hours prior tothe trip were associated with lower odds of helmet use.Personal characteristics positively associated with helmetuse included female sex, older age, higher income andhigher education.DiscussionThe most commonly used safety equipment was helmets(69% overall), even in Toronto where use of helmets isnot required of adults. This reflects the emphasis on hel-mets as “the major safety measure for bicyclists” inCanada [27]. Use of lights was uncommon (~20%), but itis required at night in both jurisdictions, and these lawswere followed by about the same proportion of cyclistsas complied with helmet legislation in Vancouver(~77%). Use of lights at dusk and dawn is also mandatedby legislation, but this was much less prevalent in ourstudy. The use of lights in the daytime was rare. Since1990, the Canadian Motor Vehicle Safety Standard hasrequired that all motor vehicles be equipped with frontdaytime running lights, so it is interesting that the po-tential for increasing the visibility of cyclists via use oflights in daytime has not been recognized either in lawor by individuals. A new development related to thisissue is bike share systems. These are being implementedTable 3 Associations between whether at least one bike light was turned on and trip or personal characteristics,variables retained in adjusted analysis only, each variable on its own (unadjusted) and in multiple logistic regression(adjusted)At least one bike light turned onYes/No (% Yes/% No)* Unadjusted Odds Ratio(95% Confidence Interval)Adjusted Odds Ratio(95% Confidence Interval)Trip time of dayDay 32/503 (6/94) 1 (reference) 1 (reference)Dawn or dusk 22/28 (44/56) 12.6 (6.49 – 24.6) 13.2 (6.42 – 27.2)Night 81/24 (77/23) 50.2 (30.0 – 90.2) 71.1 (35.8 – 141)Trip weather typeClear sky 63/414 (13/87) 1 (reference) 1 (reference)Cloud cover 39/93 (30/70) 2.91 (1.83 – 4.61) 3.43 (1.83 – 6.41)Fog, mist, rain or snow 25/35 (42/58) 4.85 (2.72 – 8.65) 3.07 (1.33 – 7.10)Wind 3/11 (21/79) 2.03 (0.55 – 7.61) 1.75 (0.30 – 10.4)Trip day of weekWeekday 109/422 (21/79) 1 (reference) 1 (reference)Weekend 26/133 (16/84) 0.73 (0.45 – 1.19) 0.48 (0.24 – 0.97)SexMale 92/318 (22/78) 1 (reference) 1 (reference)Female 43/237 (15/85) 0.60 (0.40 – 0.91) 0.56 (0.32 – 0.99)Helmet worn during tripNo 33/179 (16/84) 1 (reference) 1 (reference)Yes 102/376 (21/79) 1.45 (0.93 – 2.25) 3.15 (1.61 – 6.16)Significant associations in bold.* Row percent.Teschke et al. BMC Public Health 2012, 12:765 Page 4 of 9http://www.biomedcentral.com/1471-2458/12/765in various Canadian cities and, to date, all have bikesequipped with front and rear LED lights that are onwhenever the bicycle is moving. Bike share systems werenot in place in Toronto or Vancouver at the time of ourstudy. Use of conspicuous colours (white, yellow, orangeor red) on the torso (33%) was more common than useof lights in daytime, however the majority of participantswore other colours. Poor weather was associated withless use of conspicuous clothing, opposite to what wouldbe desirable, perhaps indicative of the typical colours ofcoats sold for cold or rainy weather. It is possible thatsome of the dark or muted coloured coats had reflectivetape that would be visible when illuminated at night, butwe did not document this in a systematic way. Brightlycoloured jackets are sold in bicycle shops, and these maybe more often purchased by frequent cyclists; they weremore likely to wear conspicuous clothing in this study.Of the three types of safety equipment examined here,helmets have been the most frequently studied. Studiesthat elicited self-reported regular use of helmets inOregon and New Zealand indicated very high propor-tions (95% or more) [19,21], but these levels seem un-realistically high compared to observations of cyclists inthe field and self-reports about a specific trip (e.g., an in-jury trip, as in this study). In US and Canadian jurisdic-tions, typical proportions of adults wearing helmets havebeen in the range of 30 to 50% where there is no legalrequirement to do so [10-12,15,16,18,22,23,25], andsomewhat over 70% where legislation requires use byadults [16]. These proportions are comparable to(though slightly lower than) our findings, perhaps be-cause our sample was skewed to regular cyclists. In con-tinental Europe, helmet use rates are considerably lower,with reports of 2% in Paris [15], 12% in Germany [14],and 2-6% among pediatricians in the Netherlands [24].A UK study reported 27% of observed cyclists wore hel-mets [20]. Factors associated with not wearing a helmetare similar to many of those found in our study: alcoholuse [10-12]; younger ages [14,16,25]; lower educationand income [14,16]; and less distance or duration of cyc-ling [14,15,25]. Studies examining sex have not foundconsistent relationships [14-16,22], though in NorthAmerica (as in our study) women appear to be morelikely to use helmets [15,16].Studies of the prevalence of light use have mainly fo-cused on use at dawn, dusk and night, rather than dur-ing the day. Several have surveyed self-reported regularuse and may suffer from over-reporting: 92% indicatedback light and 87% front light use in New Zealand [21];90% back light and 83% front light use in Australia [13];96% any light use in Portland, Oregon [19]. Those doingdirect field observations have found lower proportions:50% rear light use, 48% front light use in the UK [20]; 40to 60% use at night in New Zealand [17]. One studycompared Paris and Boston and found that 47% ofcyclists used lights at night in Paris versus 15% in Boston[15]. Factors associated with light use were rarely stud-ied, but included results similar to ours: light use wasmore common among those who wore helmets [20]; andamong men, older adults, and on weekdays [15].Table 4 Associations between whether highly visible clothing was worn on the torso and trip or personalcharacteristics, variables retained in adjusted analysis only, each variable on its own (unadjusted) and in multiplelogistic regression (adjusted)Highly visible clothing worn on the torsoYes/No (% Yes/% No)* Unadjusted Odds Ratio(95% Confidence Interval)Adjusted Odds Ratio(95% Confidence Interval)Trip weather typeClear sky 172/305 (36/64) 1 (reference) 1 (reference)Cloud cover 44/88 (33/67) 0.88 (0.53 – 1.33) 0.85 (0.56 – 1.29)Fog, mist, rain or snow 10/50 (16/84) 0.36 (0.18 – 0.72) 0.33 (0.16 – 0.68)Wind 4/10 (29/71) 0.79 (0.24 – 2.60) 0.85 (0.26 – 2.86)Age19 - 29 78/185 (30/70) 1 (reference) 1 (reference)30 - 39 54/114 (32/68) 1.15 (0.75 – 1.74) 1.19 (0.78 – 1.83)40 - 49 41/76 (35/65) 1.32 (0.83 – 2.10) 1.35 (0.84 – 2.16)50 - 59 37/46 (45/55) 1.95 (1.17 – 3.25) 1.85 (1.10 – 3.10)≥ 60 19/37 (34/66) 1.25 (0.68 – 2.32) 1.27 (0.68 – 2.38)Cycling frequency (trips per year)† 164 vs. 145 1.17 (1.05 - 1.36) 1.17 (1.05 - 1.30)Significant associations in bold.* Row percent.†= continuous variable: mean trips per year for yes vs. no; odds ratio and 95% confidence interval for 52 trips per year (equivalent to cycling once per week).Teschke et al. BMC Public Health 2012, 12:765 Page 5 of 9http://www.biomedcentral.com/1471-2458/12/765Table 5 Associations between whether helmet was worn and trip or personal characteristics, variables retained inadjusted analysis only, each variable on its own (unadjusted) and in multiple logistic regression (adjusted)Helmet wornYes/No (% Yes/% No)* Unadjusted Odds Ratio(95% Confidence Interval)Adjusted Odds Ratio(95% Confidence Interval)CityVancouver 315/99 (76/24) 1 (reference) 1 (reference)Toronto 163/113 (59/41) 0.46 (0.33 – 0.64) 0.38 (0.25 – 0.57)Bike light turned on during tripNo 376/179 (68/32) 1 (reference) 1 (reference)Yes 102/33 (76/24) 1.45 (0.93 – 2.25) 2.02 (1.17 – 3.50)Trip distance< 2 km 147/102 (59/41) 1 (reference) 1 (reference)2 - < 5 km 158/63 (71/29) 1.63 (1.10 – 2.41) 1.67 (1.05 – 2.65)5 - < 10 km 106/32 (77/23) 2.14 (1.34 – 3.44) 1.67 (0.96 – 2.89)10 - < 20 km 36/12 (75/25) 1.91 (0.94 – 3.85) 1.47 (0.65 – 3.34)≥ 20 km 31/3 (91/9) 6.75 (2.01 – 22.7) 5.43 (1.42 – 20.8)Bike type used on tripMountain bike 139/67 (67/23) 1 (reference) 1 (reference)City bike 17/12 (59/41) 0.66 (0.30 – 1.46) 0.82 (0.33 – 2.05)Touring/road bike 98/46 (68/32) 1.01 (0.64 – 1.61) 1.11 (0.65 – 1.90)Racing bike 50/16 (76/24) 1.45 (0.77 – 2.73) 1.22 (0.58 – 2.58)Folding bike 7/5 (58/42) 0.65 (0.19 – 2.12) 0.84 (0.20 – 3.44)Hybrid 152/28 (84/16) 2.58 (1.55 – 4.27) 2.08 (1.18 – 3.68)Cruiser 6/17 (26/74) 0.17 (0.07 – 0.47) 0.15 (0.05 – 0.46)BMX bike 1/6 (14/86) 0.07 (0.01 – 0.66) 0.15 (0.02 – 1.40)Fixed gear 8/15 (35/65) 0.25 (0.10 – 0.61) 0.42 (0.16 – 1.16)Alcohol used in 6 hours prior to tripNo 442/175 (72/28) 1 (reference) 1 (reference)Yes 36/37 (49/51) 0.39 (0.24 – 0.64) 0.42 (0.23 – 0.78)SexMale 272/138 (66/34) 1 (reference) 1 (reference)Female 206/74 (74/26) 1.33 (0.95 – 1.87) 1.62 (1.06 – 2.48)Age19 - 29 160/103 (61/39) 1 (reference) 1 (reference)30 - 39 114/54 (68/32) 1.33 (0.88 – 2.01) 0.98 (0.59 – 1.63)40 - 49 90/27 (77/23) 2.14 (1.29 – 3.53) 1.27 (0.67 – 2.40)50 - 59 69/14 (83/17) 3.04 (1.62 – 5.68) 2.45 (1.12 – 5.35)≥ 60 44/12 (79/21) 2.24 (1.27 – 4.45) 1.35 (0.58 – 3.17)Income< $15,000 30/36 (45/55) 0.44 (0.24 – 0.81) 0.45 (0.22 – 0.92)$15,000 - 29,999 50/33 (60/40) 0.75 (0.42 – 1.33) 0.94 (0.48 – 1.83)$30,000 - 49,999 83/37 (69/31) 1.13 (0.66 – 1.92) 1.12 (0.61 – 2.07)$50,000 - 79,999 88/45 (69/31) 1 (reference) 1 (reference)$80,000 - 119,999 80/19 (81/19) 2.11 (1.14 – 3.90) 1.67 (0.84 – 3.35)≥ $120,000 97/12 (89/11) 4.00 (1.99 – 8.06) 2.28 (1.03 – 5.07)DK/Refuse 50/30 (63/37) 0.89 (0.50 – 1.61) 0.84 (0.42 – 1.71)Teschke et al. BMC Public Health 2012, 12:765 Page 6 of 9http://www.biomedcentral.com/1471-2458/12/765Few studies have examined conspicuous clothing useand those that have suggest it is less common than useof helmets or lights at night. The clothing colours andtypes studied were not always defined or similar. A studyin Alberta, Canada observed 16% of cyclists wearing yel-low, orange or red clothing on the torso, and 19% wear-ing white [18]. A study in the UK observed 10% wearingfluorescent or reflective clothing [20]. In self-reportstudies, 30% reported regular use of fluorescent coloursin New Zealand [21] and 23% reported always use inAustralia [13]. Only one study examined features asso-ciated with use and they found highly visible clothing tobe associated with helmet use [20]. In our study, con-spicuous clothing use and helmet use were notassociated.As with studies of equipment use, studies of injuryprevention related to safety equipment have focusedon helmets. Enough studies have examined the associ-ation between helmets and injuries to allow reviewsand meta-analyses. Helmets have been shown toreduce head and face injuries (and increase neckinjuries) in the event of a crash [28] and this type ofpost-crash protection has been emphasized in NorthAmerica. In contrast, there has been little researchdirectly examining the effectiveness of lights or con-spicuous clothing as a means of preventing crashes.This may be because this type of study is much moredifficult to conduct than studies of injury type and se-verity that dominate the literature on helmets. A NewZealand study [21] found that cyclists who reportedalways wearing fluorescent colours had lower risks ofcrashes and days off work. They also found a lowercrash risk among those who reported always using aback light at night. Kwan and Mapstone [29] reviewedthe literature on visibility aids for cyclists and pedes-trians. They concluded that daytime visibility improvedwith white, yellow, orange, and red materials, and thatnight-time visibility aids (lights especially, but also re-flective clothing) enhanced detection and recognitionand shortened reaction times of observers. More re-cent studies support these results [13,18]. Our studydesign did not allow analyses of the risk of crasheswith the various types of safety equipment reportedhere. We were able to examine various surrogates ofseverity (e.g., transport by ambulance, hospitalization)and found that none of these safety equipment typeswas associated with injury severity, after controllingfor factors such as route infrastructure, weather, anddemographics [30].This study had a number of limitations. It collecteddata from injured cyclists who attended a hospital emer-gency department, so may not be representative of allcyclists. The cyclists in this study, despite beingrecruited via an injury study and despite having manypotential participants who were not contactable, mirrorcharacteristics of cyclists in North America, i.e., domin-antly male, young, and educated [3,31]. An exception isthat the sample included mainly frequent cyclists, likelya reflection of the fact that more time spent cyclingoffers greater opportunity to be injured. The study usedself-reports by injured cyclists about safety equipmentuse. Our results compared favourably to studies usingobservations of cyclists in the field [10–12,15-18,20,22,23,25]. This may be because self-reportingabout a single trip (in this case, the injury trip) is wellrecalled and accurately reported. In addition, we useddeliberate question ordering and wording to reduce thechance that responses aimed at conforming to behav-ioural norms and laws. For example, we purposely askedabout helmet use indirectly by querying the colour ofthe helmet, instead of whether a helmet was worn. Otherquestions that might be affected by social desirabilitybias are those on alcohol and drug use. In our study,10% of cyclists self-reported drinking in the 6-hour timeframe prior to the trip. In studies that measured bloodalcohol levels in more severely injured adult cyclists(fatally injured or hospitalized), 10, 14 and 19% had levelsover 0.08 g/dL [11,23,32], suggesting that our mode ofTable 5 Associations between whether helmet was worn and trip or personal characteristics, variables retained inadjusted analysis only, each variable on its own (unadjusted) and in multiple logistic regression (adjusted) (Continued)EducationSome high school 7/6 (54/46) 0.43 (0.13 – 1.42) 0.42 (0.11 – 1.70)Completed high school 19/21 (48/52) 0.30 (0.15 – 0.59) 0.30 (0.13 – 0.68)Some post-secondaryeducation71/48 (60/40) 0.47 (0.29 – 0.76) 0.61 (0.35 – 1.06)Completed college/technical diploma68/57 (54/46) 0.38 (0.24 – 0.61) 0.43 (0.25 – 0.74)Completed university degree 190/59 (76/24) 1 (reference) 1 (reference)Completed graduate degree 123/21 (85/15) 1.80 (1.04 – 3.12) 1.29 (0.70 – 2.40)Significant associations in bold.* Row percent.Teschke et al. BMC Public Health 2012, 12:765 Page 7 of 9http://www.biomedcentral.com/1471-2458/12/765questioning may have produced reasonable results for al-cohol use as well. We did not collect data on the reasonswhy safety equipment was used or not, but we were ableto examine an extensive list of personal and trip charac-teristics associated with equipment use. Finally, we didnot collect information about many other types ofindividual-based equipment that may prevent injuries tocyclists, including reflective tape, reflectors, rear-viewmirrors, disc versus rim brakes, and bells.ConclusionsIn this study of injured cyclists in two of Canada’s largestcities, we examined three types of individual-based safetyequipment and found that helmets were the most fre-quently used, and were the only type of equipment usedon the majority of trips. Helmets are a post-crash injurymitigation measure, whereas visibility aids are meant toallow other route users to detect and avoid a cyclist andthus prevent crashes from occurring. Studies of the in-jury reduction effectiveness of these pre-crash primaryprevention devices are promising but rare, so this is anarea worthy of further study. In the meantime, there isroom for increasing awareness among cyclists and cyc-ling stakeholders of the enhanced detection provided byvisibility aids and their potential to reduce collision risk.There were groups in the cycling population who wereless likely to use each type of safety equipment, suggest-ing areas of focus for change. People who tended not touse bike lights or conspicuous clothing were those whocycle less and may have less knowledge about cyclingequipment (e.g., weekend cyclists, less frequent cyclists).This suggests the potential value of communicationcampaigns like the ones that have increased helmet use.Another approach could include changes to bicyclesales, so that all commuter bikes are sold with lights (asmotor vehicles are). The population not wearing helmetsis smaller and may be more difficult to reach with add-itional messaging: those associated with risk-taking be-haviour (youth, men, people who have consumedalcohol); and those with less income and education.Strategies to prevent injuries in such populations mayrequire a different focus: safety improvements in thecycling environment (e.g., lower motor vehicle speedlimits on residential streets, dedicated bicycle infrastruc-ture including cycle tracks, bike lanes and paths)[8,26,33]. Such population-level approaches to injuryprevention benefit all cyclists and may benefit other roadusers as well.Competing interestsCCOR, MW, and PAC have held consultancies to related to theirtransportation or injury biomechanics expertise. PAC has stock in a companydeveloping a helmet that he co-invented. All other authors have no financialor other relationships or activities that could appear to have influenced thesubmitted work.Authors’ contributionsKT, MAH, CCOR, and PC were responsible for initial conception and design ofthe study. KT, MAH, CCOR, PC, MW, MC, MDC, JB, GH, SB and SMF wereresponsible for the funding proposal. MAH, CCOR, MW, MM, MDC and KTdesigned and tested data collection instruments. JB, GH, SMF, and MDCcontributed to identification of injured cyclists at the study hospitals. HS wasresponsible for data analyses. KT drafted the article. All authors contributedto study design and implementation, analysis decisions, interpretation ofresults, and critical revision of the article.AcknowledgementsWe thank the study participants for generously giving their time. Weappreciate the many contributions of study staff (Lee Vernich, Evan Beaupré,Niki Blakely, Jill Dalton, Vartouji Jazmaji, Martin Kang, Kevin McCurley, AndrewThomas), hospital personnel (Barb Boychuk, Jan Buchanan, Doug Chisholm,Nada Elfeki, Kishore Mulpuri), city personnel (Peter Stary, David Tomlinson,Barbara Wentworth) and community collaborators (Jack Becker, BonnieFenton, David Hay, Fred Sztabinski). The study was funded by the Heart andStroke Foundation of Canada and the Canadian Institutes of Health Research(Institute of Musculoskeletal Health and Arthritis, and Institute of Nutrition,Metabolism and Diabetes). JRB, MAH, and MW were supported by awardsfrom the Michael Smith Foundation for Health Research. MAH, CCOR, andMW were supported by awards from the Canadian Institutes of HealthResearch.Author details1School of Population and Public Health, 2206 East Mall, University of BritishColumbia, Vancouver, BC, Canada. 2Department of Emergency Medicine,University of British Columbia, Vancouver, BC, Canada. 3Emergency Medicine,University Health Network, Toronto, ON, Canada. 4Department of MechanicalEngineering, ICORD and the Brain Research Centre, University of BritishColumbia, Vancouver, BC, Canada. 5Occupational Cancer Research Centre,Toronto, ON, Canada. 6Liu Institute for Global Issues, University of BritishColumbia, Vancouver, BC, Canada. 7School of Public Health, University ofToronto, Toronto, ON, Canada. 8Toronto Centre for Active Transportation,Toronto, ON, Canada. 9British Columbia Injury Research and Prevention Unit,Vancouver, BC, Canada. 10Faculty of Health Sciences, Simon Fraser University,Burnaby, BC, Canada.Received: 6 June 2012 Accepted: 3 September 2012Published: 11 September 2012References1. Ogilvie D, Egan M, Hamilton V, Petticrew M: Promoting walking andcycling as an alternative to using cars: Systematic review. BMJ 2004,329:763.2. Noland RB: Perceived risk and modal choice: Risk compensation intransportation systems. Accid Anal Prev 1995, 27:503–521.3. Winters M, Davidson G, Kao D, Teschke K: Motivators and deterrents ofbicycling: comparing influences on decisions to ride. Transportation 2011,38:153–168.4. Beck LF, Dellinger AM, O'Neil ME: Motor vehicle crash injury rates bymode of travel, United States: using exposure-based methods toquantify differences. Am J Epidemiol 2007, 166:212–218.5. Transport Canada: Canadian Motor Vehicle Traffic Collision Statistics; 2009. 2011,TP 3322 http://www.tc.gc.ca/eng/roadsafety/tp-tp3322-2009-1173.htm.6. Agran PF, Castillo DN, Winn DG: Limitations of data compiled from policereports on pediatric pedestrian and bicycle motor vehicle events. AccidAnal Prev 1990, 22:361–370.7. Haddon W: On the escape of tigers: an ecologic note. Am J Public Health1970, 60:2229–2234.8. Pucher J, Buehler R: Making cycling irresistible: Lessons from theNetherlands, Denmark and Germany. Transp Rev 2008, 28:495–528.9. Reynolds CCO, Harris MA, Teschke K, Cripton PA, Winters M: The impact oftransportation infrastructure on bicycling injuries and crashes: A reviewof the literature. Environ Health 2009, 8:47.10. Crocker P, Zad O, Milling T, Lawson KA: Alcohol, bicycling, and head andbrain injury: a study of impaired cyclists' riding patterns. Am J Emerg Med2010, 28:68–72.11. Li G, Baker SP, Smialek JE, Soderstrom CA: Use of alcohol as a risk factorfor bicycling injury. JAMA 2001, 285:893–896.Teschke et al. BMC Public Health 2012, 12:765 Page 8 of 9http://www.biomedcentral.com/1471-2458/12/76512. Spaite DW, Criss EA, Weist DJ, Valenzuala TD, Judkins D, Meislin HW: Aprospective investigation of the impact of alcohol consumption onhelmet use, injury severity, medical resource utilization, and health carecosts in bicycle-related trauma. J Trauma 1995, 38:287–290.13. Wood JM, Lachereza PF, Marszaleka RP, King MJ: Drivers’ and cyclists’experiences of sharing the road: Incidents, attitudes and perceptions ofvisibility. Accid Anal Prev 2009, 41:772–776.14. Ritter N, Vance C: The determinants of bicycle helmet use: Evidence fromGermany. Accid Anal Prev 2011, 43:95–100.15. Osberg JS, Stiles SC, Asare OK: Bicycle safety behavior in Paris and Boston.Accid Anal Prev 1998, 30:679–687.16. Dennis J, Potter B, Ramsay T, Zarychanski R: The effects of provincialhelmet legislation on helmet use and bicycle ridership in Canada. InjuryPrev 2010, 16:219–224.17. Ferguson B, Blampied NM: Unenlightened: An unsuccessful attempt topromote use of cycle lights at night. Accid Anal Prev 1991, 23:561–571.18. Hagel BE, Lamy A, Rizkallah JW, Belton KL, Jhangri GS, Cherry N, Rowe BH:The prevalence and reliability of visibility aid and other risk factor datafor uninjured cyclists and pedestrians in Edmonton, Alberta, Canada.Accid Anal Prev 2007, 39:284–289.19. Hoffman MR, Lambert WE, Peck EG, Mayberry JC: Bicycle commuter injuryprevention: It is time to focus on the environment. J Trauma Inj InfectCritical Care 2010, 69:1112–1119.20. McGuire L, Smith N: Cycling safety: injury prevention in Oxford cyclists.Injury Prev 2000, 6:285–287.21. Thornley SJ, Woodward A, Langley JD, Ameratunga SN, Rodgers A:Conspicuity and bicycle crashes: preliminary findings of the TaupoBicycle Study. Injury Prev 2008, 14:11–18.22. Karkhaneh M, Rowea BH, Saundersa LD, Voaklander DC, Hagel BE: Bicyclehelmet use four years after the introduction of helmet legislation inAlberta, Canada. Accid Anal Prev 2011, 43:788–796.23. Rosenkranz KM, Sheridan RL: Trauma to adult bicyclists: a growingproblem in the urban environment. Injury 2003, 34:825–829.24. Villamor E, Hammer S, Martinez-Olaizola A: Barriers to bicycle helmet useamong Dutch paediatricians. Child Care Health Dev 2008, 34:743–747.25. Finnoff JT, Laskowski ER, Altman KL, Diehl NN: Barriers to bicycle helmetuse. Pediatrics 2001, 108:E4.26. Teschke K, Harris MA, Reynolds CCO, Winters M, Babul S, Chipman M,Cusimano MD, Brubacher J, Friedman SM, Hunte G, Monro M, Shen H,Vernich L, Cripton PA: Route infrastructure and the risk of injuries tobicyclists: A case-crossover study. Am J Public Health 2012, in press.27. Transport Canada: Road Safety in Canada.; 2011. TP 15145 E. [http://www.tc.gc.ca/eng/roadsafety/tp-tp15145-1201.htm]28. Elvik R: Publication bias and time-trend bias in meta-analysis of bicyclehelmet efficacy: a re-analysis of Attewell, Glase and McFadden, 2001.Accid Anal Prev 2011, 43:1245–1251.29. Kwan I, Mapstone J: Visibility aids for pedestrians and cyclists: asystematic review of randomised controlled trials. Accid Anal Prev 2004,36:305–312.30. Cripton P, Harris MA, Reynolds CCO, Winters M, Babul S, Chipman M,Cusimano MD, Brubacher J, Friedman SM, Hunte G, Shen H, Teschke K:Bicyclists’ injuries: Route, personal and trip factors related to severity. Inpreparation.31. Pucher J, Buehler R, Merom D, Bauman A: Walking and cycling in theUnited States, 2001–2009: Evidence from the National Household TravelSurveys. Am J Public Health 2011, 101(S1):S310–S317.32. Rowe BH, Rowe AM, Bota GW: Bicyclist and environmental factorsassociated with fatal bicycle-related trauma in Ontario. CMAJ 1995,152:45–53.33. Harris MA, Reynolds CCO, Winters M, Cripton PA, Shen H, Chipman M,Cusimano MD, Babul S, Brubacher J, Friedman SM, Hunte G, Monro M,Vernich L, Teschke K: Bicyclists’ Injuries and the Cycling Environment:Comparing the effects of infrastructure on cycling safety at intersectionsand non-intersections. Inj Prev 2012, submitted.doi:10.1186/1471-2458-12-765Cite this article as: Teschke et al.: Personal and trip characteristicsassociated with safety equipment use by injured adult bicyclists: across-sectional study. BMC Public Health 2012 12:765.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistributionSubmit your manuscript at www.biomedcentral.com/submitTeschke et al. BMC Public Health 2012, 12:765 Page 9 of 9http://www.biomedcentral.com/1471-2458/12/765


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items