What types of injuries did seriously injured pedestrians and cyclists receive in a Swedish urban region in the time period 2003–2017 when Vision Zero was implemented?

(1)

Original Research

What types of injuries did seriously injured pedestrians and cyclists

receive in a Swedish urban region in the time period 2003

e2017

when Vision Zero was implemented?

A. V€arnild

a,*

_{, P. Tillgren}

a

, P. Larm

a,b a_{School of Health, Care and Social Welfare, M€alardalen University, Sweden} b_{Department of Clinical Neuroscience, Karolinska Institutet, Sweden}

a r t i c l e i n f o

Article history:

Received 16 August 2019 Received in revised form 18 November 2019 Accepted 23 November 2019 Keywords: Road injuries Vision zero Pedestrian Cyclist Urban ISS

a b s t r a c t

Objectives: The aim of the study is to examine what types of injuries that seriously injured pedestrians and cyclists received in urban road spaces from 2003 to 2017 in the Swedish region of V€astmanland, when the road safety policy Vision Zero was implemented.

Study design: This is a cross-sectional data annually collected over a period ofﬁfteen years.

Methods: Data from health care for 403 seriously injured pedestrians and cyclists were retrieved from the registry STRADA (Swedish Trafﬁc Accident Data Acquisition) and cross-referenced with the National Road Database to see if any Vision Zero measures had previously been implemented at the crash location. The study includes injuries from both single and multiple crashes on roads, pavements, and tracks for walking and cycling (road space). Statistical analysis was performed by descriptive statistics, chi-square tests, and multiple logistic regression analyses.

Results: Pedestrians were seriously injured in lower extremities more than cyclists, whereas more cy-clists were seriously injured in the head. During the period, pedestriansꞌ head injuries decreased signiﬁcantly, but injuries in lower extremities increased signiﬁcantly. In addition, for cyclists, there was a shift from decreased probability of head injuries to increased probability of injuries in lower extremities related to increased age. For pedestrians, pavements/tracks were associated with a decreased probability of a majority of injury outcomes but for cyclists only for severe injury outcomes.

Conclusions: From 2003 to 2017, there was a shift among seriously injured pedestrians, with head in-juries decreasing and inin-juries in lower extremities increasing. This shift was probably related to an ageing population in the region, given that increased age among both pedestrians and cyclists was associated with a decreased probability of head injuries but increased probability of injuries in lower extremities. On Vision Zero roads, there was a decreased probability of pedestrians receiving serious injury to more than one bodily region. An increased number of older people combined with policies for more active mobility such as walking and cycling are a challenge for road authorities in urban areas. © 2019 The Authors. Published by Elsevier Ltd on behalf of The Royal Society for Public Health. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Introduction

Globally, 50 million people are injured annually in road trafﬁc and many of them get lifelong disabilities. The costs for injuries in most countries amount to 3% of the gross domestic product. Within the European Union, these costs can mostly be linked to

serious injuries.1,2 Thus, road trafﬁc injuries constitute a major public health problem. In response, road safety programmes have been developed in many countries, and although their measures, priorities and implementations may differ, common features include appropriate speed limits, mandatory seat belt use and efforts to prevent alcohol and drug use in trafﬁc.1,3_{The Swedish}

road safety programme Vision Zero, with the goal of zero fatalities and seriously injured road users, was initiated in 1997 with the introduction of median barriers.4 In 2006, speed cameras were installed on some road sections.5A national speed-limit revision was begun in 2009e2010 on rural and urban roads and is still

* Corresponding author: M€alardalen University, School of Health, Care and Social Welfare, Box 883, SE-721 23, V€asterås, Sweden.

E-mail addresses: astrid.varnild@mdh.se (A. V€arnild), per.tillgren@mdh.se (P. Tillgren),peter.larm@mdh.se(P. Larm).

Contents lists available atScienceDirect

Public Health

jo u rn a l h o m e p a g e :w w w . e l s e v i e r . c o m / l o c a t e / p u h e

https://doi.org/10.1016/j.puhe.2019.11.019

0033-3506/© 2019 The Authors. Published by Elsevier Ltd on behalf of The Royal Society for Public Health. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

(2)

ongoing.6In urban areas, it was allowed from the start to reduce speed limits to 30 km/h on single streets, and from 2010 it was also allowed to reduce the speed limit in larger areas from 50 km/h to 40 or 30 km/h. Furthermore, roundabouts and adapted crossings for pedestrians and cyclists have been implemented.7e9 Because of the implemented road safety pro-grammes, fatalities have decreased since the year 2000 in Sweden and in many other countries.1,10In the EU, for example, fatalities have decreased by 20% since 2010.11

Although fatalities have decreased during Vision Zero, less is known generally about the development of serious injuries, in particular among unprotected road users.1,3,12 However, an recently published study suggests that although seriously injured car occupants decreased from 2003 to 2014, the incidence of seriously injured unprotected road users increased, especially in urban areas.13A number of reasons may have contributed to the lack of knowledge about injured unprotected road users. Thefirst reason is a lack of necessary data and knowledge about what impact road safety measures have on serious injuries in different types of crashes, especially for pedestrians and cyclists.1,10,14 Second, a disagreement of whether single pedestrian crashes should be defined as road traffic crashes.15,16_{Consequently,}

sin-gle pedestrian crashes are not included in Vision Zero targets.8,15 Third, pedestrian crashes are assumed only to occur in collisions with other road users.14 Nevertheless, approximately one-third of all road trafﬁc fatalities in urban Europe consist of pedes-trians and cyclists.17 _{Furthermore, the government in Sweden}

has expressed a need to develop transport infrastructure to stimulate more walking and cycling for better environment and public health.18 Declining physical activity in Sweden and in many other high-income countries is also a leading risk factor for physical ill health, with effects on both mental health and overall quality of life.19e21 Thus, there is a potential conﬂict between more active mobility and the increased risk of road injuries for pedestrians and cyclists.22Recent studies in Sweden show that more than 80% of the serious injuries on urban roads occurred among pedestrians and cyclists.23 The number of seriously injured pedestrians and cyclists more than doubled from 2003 to 2017, and the increasing age contributed to this increase.24 However, to further understand the implications of the in-crease in seriously injured pedestrians and cyclists in urban areas, knowledge is needed about what types of injuries have increased.

In Sweden, municipalities are the road authorities responsible for urban road infrastructure and thus they are responsible for implementing and maintaining road safety measures. A regional approach is therefore necessary when trying to understand the development of serious injuries on urban roads because the implementation of Vision Zero measures can differ between re-gions. Municipalities are also responsible for the areas alongside the road such as pavements and tracks for walking and cycling. These three parts (roads, pavements and tracks) are included in a deﬁned road space, where serious injuries are studied. The aim of the present study is to examine what types of injuries that seri-ously injured pedestrians and cyclists received in urban road spaces from 2003 to 2017 in the Swedish region of V€astmanland when the road safety policy Vision Zero was implemented. Speciﬁc questions to be investigated are as follows: Which bodily regions were affected by injuries and how severe were the injuries from 2003 to 2017? What role do demographic factors (including sex and age) and part of the road space have for the location of in-juries? What role do demographic factors and part of the road space have for severity of injuries and number of injured bodily regions?

Methods

Data about injuries among seriously injured pedestrians and cyclists Data were retrieved from the registry STRADA (Swedish Traffic Accident Data Acquisition) for afifteen-year period from 2003 to 2017.25,26STRADA contains information from emergency hospitals and police reports. From the STRADA data, seriously injured pe-destrians and cyclists in Region V€astmanland were identified in total 403 individuals (206 pedestrians [n¼ 206] and 197 cyclists). Although, STRADA has been a national registry only since 2016, the region of V€astmanland is one of two Swedish regions that have reported data since 2000.26 STRADA also includes information about where the crashes happened, and these locations (roads, pavements and tracks) were checked in the National Road Database27to see if there had been any implementation of Vision Zero measures before the crash.

Measurements

The definition of a seriously injured pedestrian or cyclist is Injury Severity Score (ISS)>8, which is the same definition as the registry applies. The ISS score is based on the Abbreviated Injury Scale (AIS) and is the sum of squared AIS scores for a maximum of three out of six defined bodily regions.28_{When the ISS score is}

calculated, only the injury with the highest AIS score in each bodily region is included. In this study, the ISS score can range from 9 to 75. Injuries in different bodily regions were then grouped into three categories based on how common they are: (1) head, face and neck; (2) lower extremities, including pelvic skeleton; and (3) thorax, abdomen, spine, upper extremities and external injuries. To mea-sure disability magnitude, pedestrians and cyclists were classiﬁed into those having one injured bodily region and those having two or three injured regions. In the same way, pedestrians and cyclists were divided between those with an ISS 9 and those above.

Factors examined for possible impact on injuries to different bodily regions, ISS score and the number of injured bodily regions include demographic factors such as sex and age, study year and road space. Age is divided into four categories as follows: 0e49 years, 50e64 years, 65e79 years and 80 years. The study years from 2003 to 2017 have been divided intoﬁve three-year periods to avoidﬂuctuations that can occur between single years. These are 2003_{e2005, 2006e2008, 2009e2011, 2012e2014 and 2015e2017.} Road space includes two indicators of importance for severe in-juries among pedestrians and cyclists: (1) where in the road space that the accident occurred, which in addition to the road also in-cludes pavements and tracks; and (2) whether or not Vision Zero measures such as reduced speed from 50 km/h to 30 or 40 km/h, safe passages by underpass or overpass, and roundabouts have been implemented.

Two covariates are included: (1) Whether the cause of the crash was related to maintenance issues or other causes; and (2) The population size of the urban area in municipalities is divided into more or less than 100,000 inhabitants.

Statistical analysis

First, descriptive statistics of how study variables are distributed in the sample and whether they differ between pedestrians and cyclists are tested with chi-square tests. Second, the distribution of injured bodily regions and severity of injuries (ISS) from 2003 to 2017 is presented separately for pedestrians and cyclists. The study period was divided into three ﬁve-year categories (2003e2007, 2008e2012 and 2013e2017) to increase statistical power owing to

(3)

low base rates. Comparisons between the two most recent periods (2008e2012 and 2013e2017) and the earliest period (2003e2007) were conducted with chi-square tests. Third, the impact of de-mographic factors (sex and age) and road space part (roads, pave-ments and tracks) on the three different bodily regions were analysed in a series of multiple logistic regression analyses that provided odds ratios (ORs) and 95% conﬁdence intervals (CIs) adjusted for causes of crashes and municipality size. In addition, ﬁve three-year study periods were included in the models to assess their impact. Analyses were conducted separately for pedestrians and cyclists. Fourth, a similar series of multiple logistic regression analyses were conducted to identify the impact of demographic factors and road space part on the severity of injuries including ISS score_{10 and severe injuries to multiple bodily regions.}

Results

Differences in the distribution of injuries, demographic factors and road space characteristics between seriously injured pedes-trians and cyclists are presented in Table 1. Over aﬁfteen-year period in Region V€astmanland, more pedestrians were seriously injured in the lower extremities, including pelvic skeleton than cyclists: 71.4% compared with 43.7%. In contrast, cyclists more frequently had head injuries (head/face/neck) and injuries to other parts of the body than pedestrians. In addition, compared with pedestrians, cyclists were more often injured in multiple bodily regions, and their injuries were more severe (ISS score10). The development of serious injuries from 2003 to 2018 among pedes-trians and cyclists is presented in Table 2. A somewhat clearer

picture emerged for pedestrians than for cyclists. Serious injuries in the bodily region of the head decreased significantly among pe-destrians from 60.7% in 2003e2007 to 24.2% in 2013e2017. In contrast, injuries in the lower extremities increased from 46.4% to 78.2% during the same period. The proportion of pedestrians with severe injuries giving an ISS score higher than 9 decreased from 46.4% to 21.8%. Furthermore, although injuries in the lower ex-tremities and thorax regions seem to have increased among cyclists during 2003e2007 and 2013e2017, and injuries in the head region seem to have decreased, thesefindings are not significant and more diverged owing to low base rates.

The influence of period categories, demographic factors and road space on body part injuries, as derived from multiple logistic regression analysis adjusted for cause of crash and urban munici-pality size, is shown inTable 3. For pedestrians, more recent time periods, older age and pavement/track were influential. The prob-ability of injuries in the head region decreased from 2012, whereas the probability of lower extremities injuries increased, in particular for the time period 2012e2014 (OR ¼ 7.87, 95% CI ¼ 1.46e42.43). The probability of injuries in the head and thorax regions decreased from age 65 with the exception of the age period80 for head injuries. In contrast, older age increased the probability of injuries in the lower extremities region. Furthermore, the probability of injuries in the thorax and lower extremities regions decreased on pavements/tracks compared with non-Vision Zero roads. For cy-clists, increased age was the most influential factor. Increased age reduced the probability of injuries in the head region but increased the probability of injuries in the lower extremities region. In addition, females had an increased probability of head region in-juries (OR¼ 2.51, 95% CI ¼ 1.19e5.30).

The influence of period categories, demographic factors and road space on injury severity and injuries in multiple body parts, as derived from multiple logistic regression analysis and adjusted for cause of crash and municipality is shown inTable 4. For pedestrians, where in the road space that the accident occurred was most influential. Accidents on pavements/tracks decreased the proba-bility of severity score higher than 9 and injuries in multiple bodily regions, whereas Vision Zero areas exhibited a decreased proba-bility of injuries in multiple bodily regions compared with non-Vision Zero areas. For cyclists, increased age was most in_fluential, with ages higher than 79 years decreasing the probability of severe ISS higher than 9 (OR¼ 0.30, 95% CI ¼ 0.10e0.90) and age higher than 64 years decreasing the probability of injuries in multiple

Table 1

Injured bodily regions (N¼ 515) and other characteristics of seriously injured cy-clists and pedestrians (N¼ 403) in Region V€astmanland 2003e2017.

Variables Pedestrians

(n¼ 206)

Cyclists (n¼ 197)

P-value

Injured bodily regions (N¼ 515)

Head, face neck 32.0% (66) 57.9% (114) <0.001 Lower extremities, pelvic skeleton 71.4% (147) 43.7% (86) <0.001 Other parts of the body 18.9% (39) 32.0% (63) 0.003 Injured bodily regions per injured person <0.001

¼ 1 75.7% (156) 54.8% (108) >1 24.3% (50) 45.2% (89) ISS scores 0.007 ¼ 9 70.0% (144) 56.9% (112) 10 30.0% (62) 43.1% (85) Sex <0.001 Men 39.3% (81) 64.0% (126) Women 60.7% (125) 36.0% (71)

Seriously injured period <0.001

2003e2005 3.9% (8) 15.7% (31) 2006e2008 12.6% (26) 21.8% (43) 2009e2011 33.5% (69) 22.9%% (45) 2012e2014 21.4% (44) 21.3% (42) 2015e2017 28.6% (59) 18.3% (36) Age category <0.001 Category 0-49 15.5% (32) 31.5% (62) Category 50-64 14.1% (29) 22.8% (45) Category 65-79 39.3% (81) 31.5% (62) Category 80 31.1% (64) 14.2% (28) Road space <0.001

Non-Vision Zero road 15.5% (32) 32.0% (63) Vision Zero road 16.5% (34) 17.2% (34) Pavement/track 68.0% (140) 50.8% (100) Cause of crash <0.001 Maintenance 64.6% (133) 21.3% (42) Other causes 35.4% (73) 78.7% (155) Municipality-urban area 100,000 inhabitants 59.7% (123) 66.0% (130) 0.192 <100,000 inhabitants 40.3% (83) 34.0% (67) ISS¼ injury severity score.

Table 2

Distribution of injured bodily regions and severity of injuries (ISS) for seriously injured pedestrians and cyclists in Region V€astmanland 2003e2007, 2008e2012 and 2013e2017.

2003e2007 2008e2012 2013e2017

Head/face/neck Pedestrians 60.7% (17) 30.8% (28)** 24.1% (21)*** Cyclists 57.6% (34) 69.7% (46) 47.2% (34) Lower extremities Pedestrians 46.4% (13) 72.5% (66)* 78.2% (68)** Cyclists 40.7% (24) 31.8% (21) 56.9% (41) Thorax/abdomen/spine/upper extremities/external injuries

Pedestrians 21.4% (6) 19.8% (18) 17.2% (15)

Cyclists 22% (13) 39.4% (26)* 33.3% (24)

ISS score 10

Pedestrians 46.4% (13) 33% (30) 21.8% (19)* Cyclists 40.7% (24) 54.5% (36) 34.7% (25) Multiple injured bodily regions

Pedestrians 35.7% (10) 25.3% (23) 19.5% (17) Cyclists 37.3% (22) 54.5% (36) 43.1% (31) Note. ***P¼ .001, **P .01, *P .05 in comparison to 2003e2007. ISS¼ Injury Severity Score

(4)

bodily regions. In addition, accidents that occurred on pavements/ tracks exhibited a decreased probability of severe ISS higher than 9 (OR¼ 0.27, 95% CI ¼ 0.13e0.60) compared with non-Vision Zero roads.

Discussion

During the 15 years when Vision Zero was implemented in the Swedish Region of V€astmanland, cyclists were more often seriously injured in the head, -a, had an ISS higher than 9 and were seriously injured in multiple body parts, whereas pedestrians more often were seriously injured in the lower extremities. Interestingly, these differences were not distinct in the earliest time period 2003e2007 and seem to have emerged thereafter with considerably reduced head injuries and injuries having an ISS higher than 9, but considerably increased injuries in the lower extremities among pedestrians. A number of factors may have contributed to this injury alleviation overtime, from head injuries to injuries in lower extremities and less severe injuries among pedestrians.

First, serious injuries among pedestrians mostly occur on pavements/tracks: 68% as opposed to 50.8% for cyclists. The multivariate analyses showed that pavements/tracks were associ-ated with a decreased probability of serious injuries to the thorax and lower extremities, an ISS score higher than 9, and serious in-juries in multiple bodily regions for pedestrians, but only for the ISS higher than 9 for cyclists. The difference between pedestrians and cyclists may reﬂect differences between access to infrastructure for their own mode of transport. Pedestrians in urban areas mostly have access to pavements and/or tracks instead of roads, whereas cyclists only have the choice between tracks and roads. However, the total national distance travelled by walking is also nearly 50% higher than that of cyclists.29Vision Zero recommends separation

Table 3

Adjusted odds ratios and 95% conﬁdence intervals for serious injuries in Region V€astmanland stratiﬁed by different parts of the body, demographic characteristics, different three-year periods and location in the road space 2003e2017.

Road user variable (reference) Head/face/neck Thorax/abdomen/spine/upper extremities/external injuries Lower extremities/pelvic skeleton

OR (95% CI) OR (95% CI) OR (95% CI)

Pedestrians

Sex (men) 0.66 (0.34e1.26) 1.84 (0.80e4.23) 1.63 (0.83e3.20)

Periods (2003e2005)

2006e2008 0.32 (0.05e1.88) 0.33 (0.05e2.20) 2.75 (0.50e15.23)

2009e2011 0.19 (0.04e1.01) 0.32 (0.06e1.80) 4.54 (0.94e22.00)

2012e2014 0.16 (0.03e0.91) 0.39 (0.07e2.33) 7.87 (1.46e42.43)

2015e2017 0.15 (0.03e0.82) 0.26 (0.04e1.56) 4.32 (0.86e21.64)

Age category (0e49 years)

50-64 years 0.54 (0.18e1.66) 0.34 (0.09e1.21) 3.70 (1.11e12.29)

65-79 years 0.36 (0.14e0.94) 0.29 (0.11e0.79) 2.94 (1.14e7.56)

80 years 0.54 (0.21e1.21) 0.30 (0.11e0.87) 3.60 (1.36e9.57)

Road space (N-VZ road)

Vision Zero road 0.46 (0.15e1.45) 0.45 (0.13e1.56) 0.37 (0.10e1.38)

Pavement/track 0.50 (0.21e1.21) 0.29 (0.11e0.76) 0.28 (0.09e0.84)

Cyclists

Sex (men) 2.51 (1.19e5.30) 0.92 (0.47e1.83) 0.65 (0.33e1.29)

Periods (2003e2005)

2006e2008 0.63 (0.20e1.97) 1.13 (0.35e3.67) 1.09 (0.37e3.21)

2009e2011 1.66 (0.49e5.56) 3.02 (0.98e9.30) 0.80 (0.26e2.47)

2012e2014 1.11 (0.35e3.46) 2.10 (0.68e6.45) 1.37 (0.47e3.98)

2015e2017 0.64 (0.20e2.10) 3.95 (1.21e12.85) 1.38 (0.45e4.20)

Age category (0e49 years)

50-64 years 0.28 (0.10e0.77) 0.98 (0.42e2.30) 3.98 (1.56e10.16)

65-79 years 0.12 (0.05e0.31) 0.47 (0.20e1.09) 7.54 (3.11e18.27)

80 years 0.09 (0.03e0.28) 0.50 (0.16e1.51) 8.30 (2.85e24.18)

Vision Zero road 1.07 (0.38e3.01) 0.40 (0.14e1.14) 1.14 (0.43e2.97)

Pavement/Track 0.45 (0.19e1.05) 0.52 (0.24e1.15) 1.15 (0.65e3.16)

Note: Adjusted for cause of crash and municipality size. VZ¼ Vision Zero; OR ¼ odds ratio; CI ¼ conﬁdence interval.

Table 4

Adjusted odds ratios and 95% conﬁdence intervals for serious injuries in Region V€astmanland stratiﬁed by ISS score, number of injured bodily regions, demographic characteristics, different three-year periods and location in the road space 2003e2017.

Road user variable (reference) ISS score10 >1 injured bodily region OR (95% CI) OR (95% CI)

Pedestrians

Sex (men) 1.13 (0.57e2.26) 1.46 (0.68e3.15) Periods (2003e2005)

2006e2008 0.63 (0.11e3.63) 0.76 (0.12e4.90) 2009e2011 0.37 (0.07e1.85) 0.36 (0.06e2.04) 2012e2014 0.27 (0.05e1.47) 0.40 (0.97e2.42) 2015e2017 0.16 (0.03e0.86) 0.21 (0.03e1.26) Age category (0e49 years)

50-64 0.37 (0.11e1.24) 0.46 (0.13e1.63) 65-79 0.52 (0.20e1.36) 0.52 (0.19e1.44)

80 0.59 (0.23e1.55) 0.71 (0.26e1.96)

Road space (Non-VZ road)

Vision Zero road 0.37 (0.12e1.13) 0.27 (0.08e0.88) Pavement/track 0.20 (0.08e0.50) 0.15 (0.06e0.40) Cyclists

Sex (men) 0.73 (0.38e1.42) 0.90 (0.48e1.72) Periods (2003e2005)

2006e2008 0.83 (0.29e2.36) 1.40 (0.49e3.98) 2009e2011 1.50 (0.53e4.32) 2.05 (0.72e5.87) 2012e2014 0.62 (0.21e1.79) 2.29 (0.81e6.47) 2015e2017 1.46 (0.48e4.43) 1.93 (0.64e5.83) Age category (0e49 years)

50-64 1.10 (0.47e2.53) 0.48 (0.21e1.11) 65-79 0.50 (0.22e1.12) 0.35 (0.16e0.76)

80 0.30 (0.10e0.90) 0.24 (0.09e0.69)

Vision Zero road 0.56 (0.21e1.45) 0.86 (0.34e2.20) Pavement/track 0.27 (0.13e0.60) 0.60 (0.28e1.28) Note: Adjusted for cause of crash and municipality size.

VZ¼ Vision Zero; OR ¼ odds ratio; CI ¼ conﬁdence interval; ISS ¼ Injury Severity Score.

(5)

of unprotected road users from motor trafﬁc,7,9_{but the supply of}

tracks in regional urban areas remained nearly the same throughout the examined period.

Second, the multivariate analyses showed that increased age among both pedestrians and cyclists was associated with a decreased probability of head injuries but increased probability of injuries in lower extremities. Thus, the shift from head injuries to lower extremity injuries among pedestrians may reﬂect the ageing population. In Europe, the average age is increasing owing to the 1940s baby boom combined with increased life expectancy.17,30,31 In the Region of V€astmanland, the population of people 65 years and older has similarly increased between 2003 and 2017.24,32An ageing population will result in 1%e3% more hip fractures annually, regardless of where the accident occurs, and until 2050, the global incidence of hip fractures is estimated to increase fourfold.33e35 There are studies suggesting that the increase in hip fractures is a result of osteoporosis and an increased fall tendency for older people, especially in the United States and the Scandinavian countries.33,36,37 For example, one Norwegian study report that 17.5% of hip fractures in people 50 years and older during a period ofﬁfteen years occurred in an outdoor transport area. The injured were considerably younger than people who got hip fractures in other areas.38

Third, Vision Zero reconstructions of urban roads were generally not associated with injury outcomes, with one exception: a decreased probability of serious injuries in multiple bodily regions for pedestrians. However, at least for pedestrians, Vision Zero roads decreased the probability of all injury outcomes compared with non-Vision Zero roads, with OR varying between 0.27 and 0.46. Only the lowest of these OR is signiﬁcant; however, they would probably become signiﬁcant with an increased sample size. Thus, it is more likely than not that speed limit reductions on urban roads from 50 km/h to 40 or 30 km/h, safe passages by means of un-derpasses or overpasses, and roundabouts9,39have reduced serious injuries in all bodily regions for pedestrians.

Implications for public health

The increasing number of older people in European countries, United States and many other countries,24,30,31together with public health policies that promote increased physical activity through walking and cycling, pose a traf_{fic-safety challenge for unprotected} road users. Pedestrians' and cyclistsꞌ injuries need to be made visible and defined as resulting from road traffic crashes to facilitate the work to make cities safe for all vulnerable road users in accordance with Agenda 2030.40

Strengths and limitations

The strengths of the study are the quality of data from health care for aﬁfteen-year period and the comparison between seri-ously injured pedestrians and cyclists in both single and multiple crashes because pedestrians in single crashes generally are excluded. The examined crashes take place in a de_{ﬁned road} space, where only one road authority is responsible for imple-menting new road safety measures. With more knowledge about where crashes take place, it is possible to decide which parts of the road space to prioritise when new measures are to be implemented.3,39

There are statistical limitations of the study, despite the extended length of the examined period. A longer period or a larger population would have given more statistical power and increased opportunities to study more factors impacting on serious injuries.

Conclusion

Three main conclusions can be drawn from this study. First, during the period, there was a shift among seriously injured pe-destrians, with head injuries decreasing and injuries of lower ex-tremities increasing. This shift is probably related to the increased average age of the people in the region because higher age was associated among both pedestrians and cyclists with a decreased probability of head injuries and an increased probability of injuries in the lower extremities. Second, pavements/tracks were associated with a decreased probability for the majority of injury outcomes among pedestrians but only for really severe injuries among cy-clists. Third, there was a decreased probability of pedestrians being seriously injured in more than one bodily region on roads trans-formed by Vision Zero.

Author statements Ethics approval

The study has been approved by the Swedish Ethical Review Authority (Dnr 2019-00209).

Funding

This study was supported by the Swedish Transport Adminis-tration (TRV 2014/49856).

Competing interests None declared. Author contributions

A.V. was the principal investigator and collector of data. P.L. and A.V. were responsible for statistical analysis. A.V. drafted the manuscript, which was discussed and reviewed by P.L. and P.T. All authors discussed the design and approved theﬁnal manuscript. References

1. International Trafﬁc Safety Data and Analysis Group (IRTAD). Road safety annual report 2018. Paris [France]: OECD/ITF; 2018.

2. World Health Organization. Road trafﬁc injuries. Fact Sheet; 2018.https://www. who.int/news-room/fact-sheets/detail/road-trafﬁc-injuries. [Accessed 13 August 2019].

3. World Health Organization. Global status report on road safety 2018. Geneva [Switzerland]: WHO Press; 2018.

4. Bergh T, Remgård M, Carlsson A, Olstam J, Str€omgren P. 2þ1-roads recent Swedish capacity and level-of-service experience. Transportation Research Procedia 2016;15:331e5.https://doi:10.1016/j.trpro.2016.06.028.

5. Belin MÅ, Tillgren P, Vedung E, Cameron M, Tingvall C. Speed cameras in Sweden and Victoria, Australia e a case study. Accid Anal Prev 2010;42: 2165e70.https://doi:10.1016/j.aap.2010.07.010.

6. Vadeby A, Forsman Å. Trafﬁc safety effects of new speed limits in Sweden. Accid Anal Prev 2018;114:34e9. https://doi.org/10.1016/j.aap.2017.02.0030001-4575.

7. Johansson R. Vision Zeroe implementing a policy for trafﬁc safety. Saf Sci 2009;47(6):826e31.https://doi:10.1016/j.ssci.2008.10.023.

8. Swedish Transport Administration. Analys av traﬁks€akerhetsutvecklingen 2018. Målstyrning av traﬁks€akerhetsarbetet mot etappmålen 2020 [Analysis of road safety developments 2018. Management by objectives of road safety work to achieve the 2020 interim targets]. Borl€ange [Sweden]: Swedish Transport Administration; 2019.

9. Tingvall C, Haworth N. Vision Zeroe an ethical approach to safety and mobility. In: Paper presented at the 6th ITE international conference road safety& trafﬁc enforcement: beyond 2000, Melbourne [Australia]; 1999.

10. Wegman F, Berg HY, Cameron I, Thompson C, Siegrist S, Weijemars W. Evi-dence-based and data-driven road safety management. IATSS Res 2015;39: 19e25.https://doi.org/10.1016/j.iatssr.2015.04.0010386-1112.

11. European Commission. Mobility and transport. Road Safety; 2018.http://ec. europa.eu/roadsafety. [Accessed 13 August 2019].

(6)

12. Weijermars W, Bos N, Filtness A, Brown L, Bauer R, Dupont E, et al. Burden of injury of serious road injuries in six EU countries. Accid Anal Prev 2018;111: 184e92.https://doi.org/10.1016/j.aap.2017.11.040.

13. V€arnild A, Larm P, Tillgren P. Incidence of seriously injured road users in a Swedish region, 2003e2014, from the perspective of a national road safety policy. 2019. BMC Public Health 19(1): 1576.

14. Aarts LT, Commandeur JJF, Welsh R, Niesen S, Lerner M, Thomas P, et al. Study on serious road trafﬁc injuries in the EU. Brussels [Belgium]: European Com-mission; 2016.

15. Berg HY, Ifver J, Hasselberg M. Public health consequences of road trafﬁc in-juriese estimation of seriously injured persons based on risk for permanent medical impairment. Transp Res Part F 2016;38:1e6.https://doi.org/10.1016/ j.trf.2015.12.007.

16. Methorst R, Schepers P, Christie N, Dijst M, Risser R, Sauter D, et al. 'Pedestrian falls' as necessary addition to the current deﬁnition of trafﬁc crashes for improved public health policies. J Transport Health 2017;6:10e2. https:// doi.org/10.1016/j.jth.2017.02.005.

17. Methorst R, Eenik R, Cardoso J, Machata K, Malsek J. Single unprotected road user crashes: Europe we have a problem. Transport Res Procedia 2016;14: 2297e305.https://doi:10.1016/j.trpro.2016.05.246.

18. Swedish Government. Uppdrag att ta fram f€orslag till nationell fikslags€overgripande plan f€or utveckling av transportsystemet och tra-fikslags€overgripande l€ansplaner f€or regional transportinfrastruktur [Mandate to develop proposals for national transport master plan for development of transport and intermodal county plans for regional transport infrastructure]. Stockholm [Sweden]: Government officies; 2012 (Government Letter 2012/ 13:119).

19. Baum F. The new public health. 4th ed. Melbourne [Australia]: Oxford University Press; 2016.

20. Guthold R, Stevens GA, Riley LM, Bull FC. Worldwide trends in insufﬁcient physical activity from 2001 to 2016: a pooled analysis of 358 population-based surveys with 1.9 million participants. Lancet Glob Health 2018;6:e1077e86. https://doi.org/10.1016/S2214-109X(18)30357-7.

21. Swedish government agency for transport policy analysis. Uppf€oljning av de transportpolitiska målen 2018 [Follow-up of the Transport Policy Objectives 2018]. Stockholm [Sweden]: Swedish Government Agency for Transport Policy Analysis; 2018. 2018.

22. Mueller N, Rojas-Rueda D, Cole-Hunter T, de Nazelle A, Dons E, Gerike R, et al. Health impact assessment of active transportation: a systematic review. Prev Med 2015;76:103e14.https://doi.org/10.1016/j.ypmed.2015.04.010. 23. V€arnild A, Belin MÅ, Tillgren P. VISION zero e road trafﬁc effects for severely

injured IN a Swedish county. In: Paper presented at the safety 2016 world conference. Finland: Tammerfors; 2016.

24. V€arnild A, Tillgren P, Larm P. Factors related to the increasing number of seriously injured cyclists and pedestrians in a Swedish urban region 2003-17. J Public Health 2019;41.https://doi.org/10.1093/pubmed/fdz064. In press.

25. Swedish Transport Agency. STRADA (Swedish trafﬁc accident data acquisition). https://transportstyrelsen.se/sv/vagtraﬁk/statistik/Olycksstatistik/om-strada. [Accessed 13 August 2019].

26.Fredlund T, Frank M. The Swedish national information system for trafﬁc ac-cidents and injuries. In: Paper presented at the safety 2016 world conference. Finland: Tammerfors; 2016.

27. Swedish Transport Administration. NVDB (National Road Database).https:// www.traﬁkverket.se/tjanster/system-och-verktyg/data/Nationell-vagdatabas. [Accessed 13 August 2019].

28. Genarelli TA, Wodzin E. AIS 2005: a contemporary injury scale. Injury Int J Care Inju 2006;37:1083e91.https://doi.org/10.1016/j.injury.2006.07.009. 29.Swedish government agency for transport policy analysis. Cyklandets utveckling

i Sverige 1995-2014e en analys av de nationella resvaneunders€okningarna. [The development of cycling in Sweden 1995e2014 e an analysis of the national travel surveys]. Stockholm [Sweden]: Swedish government agency for transport policy analysis; 2015.

30. Christensen K, Doblhammer G, Rau R, Vaupel JW. Ageing populations: the challenges ahead. Lancet 2009;374:1196e208. https://doi:10.1016/S0140-6736(09)61460-4.

31. Rechel B, Grundy E, Robine JM, Cylus J, Mackenbach JP, Knai, et al. Ageing in the European union. Lancet 2013;381:1312e22. https://doi.org/10.1016/S0140-6736(12)62087-x.

32. Statistics Sweden. Population.http://www.statistikdatabasen.scb.se/pxweb/sv/ ssd/START__BE__BE0101__BE0101A/BefolkningNy/table/tableViewLayout1/? rxid¼cf98fcf4-052e-48b3-9cef-9fdf611ed492. [Accessed 13 August 2019]. 33. Cummings SR, Melton LJ. Epidemiology and outcomes of osteoporotic fractures.

Lancet 2002;359:1761e7.https://doi.org/10.1016/S0140-6736(02)08657-9. 34.Gullberg B, Johnell O, Kanis JA. World-wide projections for hip fracture.

Osteoporos Int 1997;7(5):407e13.

35. Robinovitch SN, Evans SL, Minns J, Laing AC, Kannus P, Cripton PA, et al. Hip protectors: recommendations for biomechanical testing-an international consensus statement (part I). Osteoporos Int 2009;20(12):1977e88. https:// doi.org/10.1007/s00198-009-1045-4.

36.Carlsson A, Svensson M. Fallolyckor bland gångtraﬁkanter [Fall accidents among pedestrians, TRV 2013/71828]. G€oteborg [Sweden]: Chalmers University of Technology; 2015.

37.L€ofman O. Epidemiologin f€or frakturer [epidemiology of fractures]. Lakartid-ningen 2006;40:2956e8.

38. Emaus N, Olsen R, Ahmed LA, Balteskard L, Jacobsen BK, Magnus T, et al. Hip fractures in a city in Northern Norway over 15 years: time trends, seasonal variation and mortality. The Harstad Injury Prevention Study. Osteoporos Int 2011;22:2603e10.https://doi.org/10.1007/s00198-010-1485-x.

39.Elvik R, Høye A, Vaa T, Sørensen M. The handbook of road safety measures. United Kingdom: Emerald Group Publishing Limited; 2009.

40.United Nations General Assembly. Transforming our world: the 2030 Agenda for sustainable development (A/RES/70/1. New York [US]: United Nations; 2015.