Road safety effectiveness of Safe Route to Schools in Barcelona

Initiative details

Safe Routes to School Program in Barcelona
The "Camí escolar, espai amic" (“Safe Route to School, friendly space”) program began with the aim of increasing children´s and adolescents personal autonomy, responsibility and quality of life on their way to school or while walking around the neighborhood.

The program promotes road safety education in schools through an educational program conducted within the school and the community, and changes in the environment around the school. It is led by the Municipal Institute of Education of Barcelona (IMEB) and carried out in collaboration with the Department of Safety & Mobility of Barcelona. In each education center it includes four phases: Phase 0: We start to walk, for the definition of the project; Phase I: We explore the path, to carry out the diagnosis. Phase II. We go out into the neighborhood, to create the network of friendly spaces and celebrate the work done. Phase III: We keep the path alive, to evaluate and guarantee the sustainability of the project in the school and the educational community.

In Barcelona, police data on traffic injuries is geocoded and allows time series analyses since 2002. There are around 9,000 injuries due to road traffic crashes each year, with 11,000 persons injured, around 300 severely injured, and 15-25 fatalities in the most recent years. Pedestrians represent around 13% of all casualties, with the highest incidence occurring among young and old people.

This study aimed to evaluate the effectiveness of the Safe Routes to School (SRTS) program “Camins Escolars” conducted in Barcelona from 2006 to 2016 in reducing the number of road traffic collisions and injuries in the school environment.

The study used a pre-post quasi-experimental evaluation design, with a matched comparison group, including 64 intervention schools and 63 comparison schools. Outcomes included collisions and people injured within a radius of about 200 meters around schools during school hours, using geocoded data (2002-2019) from the local police register, and contextual variables. For each outcome measure, we fitted generalized linear mixed model with Poisson distribution.

Initiative date

to

Who was/is your target audience?

Policy makers
Public authorities

Topic

Create awareness
Provide alternative solutions

Organisation details

Road safety effectiveness of safe route to schools in Barcelona
Local Public authority
Spain
Barcelona

Contact name

Catherine Pérez

Telephone number

932027788

Project activities

If you work together with external partners, list the most important partners and briefly describe their role.

This work has been done in collaboration with other sectors:
• Barcelona Local Police, which provided information on road traffic crashes and injuries.
• The Municipal Institute on Education of Barcelona (IMEB), who was responsible of promoting and leading the project of Safe Routes To School in Barcelona. Provided detailed information on the intervention schools.
• The Municipal Institute of Informatics (IMI) provided a map with the geolocation of all the schools in the city, which allowed identification of the location of the intervention and comparison schools.
• The Department of Safety & Mobility of Barcelona, provided a map of all the street sections in the city with information on the location of points for measuring traffic congestion in the city and kilometers of the street

Please describe the project activities you carried/are carrying out and the time period over which these were implemented.

In this section we describe with detail the methods and main results.

Methods:
This was a pre-post quasi-experimental evaluation study, with a matched comparison group. The study population consisted of people who moved around Barcelona between 2002 and 2019 in Barcelona city.

The intervention group consisted of schools in which the SRTS program was conducted, and the comparison group comprised schools without the SRTS program.

The study area was defined as a buffer around the schools within a radius of about 200 meters.

The inclusion and exclusion criteria were as follows: of the 152 schools in Barcelona in which the SRTS program was implemented, we selected schools whose SRTS implementation year started after 2005 and whose inauguration (street works and program implemented) year was prior to 2016. Therefore, all the selected intervened schools had a pre-intervention and post-intervention period of at least 4 years per period.

For each intervention school, we included a close non-intervention school as a comparison school, preferably in the same neighborhood, with the same level of education (nursery, infant, primary, and secondary- or high- schools). We excluded comparison schools overlapping with the study area of an intervention school.

Finally, 64 intervention schools fulfilled the criteria, with a matched comparison school meeting the inclusion criteria for 63 of them.

Information Sources:
Barcelona Local Police Register of Road Traffic Accidents and Victims, which provided information on traffic collisions in the city from 2002 to 2019. Data on road traffic collisions and people injured included geocodes and allowed identification of collisions occurring in the intervention and comparison study zones.

The IMEB provided information on the intervention schools. For each calendar year, the IMEB defined the schools that had started the activities or interventions related to the program in that year (start year), as well as those that had carried out most of the agreed actions on the public road and their SRTS program had already been inaugurated (inauguration year).

The Municipal Institute of Informatics (IMI) provided a map with the geolocation of all the schools in the city, which allowed identification of the location of the intervention and comparison schools.

The Department of Safety & Mobility of Barcelona, provided a map of all the street sections in the city with information on the location of points for measuring traffic congestion in the city and kilometers of the street.

Study period:
The study period comprised 2002 to 2019. Three periods were defined, specifics for each school:
First, pre-intervention period, from 2002 to the year before the program started; second implementation period, from the program (works) start year to the inauguration year and third post-intervention period, from the year following the inauguration year until 2019.
Not all schools started the program in the same year, nor did they inaugurate in the same year. Therefore, not all schools had the same pre- and post- period, nor did the periods have the same number of years, although they all had a minimum of 4 pre- and 4 post- years. The two periods analyzed were the pre- and post-intervention periods.

Inclusion and exclusion criteria for traffic collisions
Traffic collisions with casualties occurring in the study area (buffer with a 200 meters radius around the schools): from Monday to Friday, from 7:00 to 9:30, 12:00-15:00 and 16:00 to 18:00, and from September 15 to June 23, were included.
Collisions occurring on Christmas and Easter holiday periods were excluded.

Study variables
Dependent or outcome variables:
Number of road traffic collisions with casualties
Number of collisions involving 0-16 years-old injured
Number of collisions involving 0-16 years-old pedestrians injured

Number of people injured
Number of 0-16 years-old injured
Number of 0-16 years-old pedestrians injured

Exposure variables:
Number of students at the school in 2018
Total population and population aged 0-16 years old resident in the school neighborhood for each study year (2002-2019)
Family Available Income in the school neighborhood in 2017
Number of injured people and number of injured pedestrians per 100 meter of street in the school neighborhood in 2018
Number of injured people and number of injured pedestrians on weekdays per 10 million km traveled by motor vehicles in the school neighborhood in 2018
Kilometer of street and kilometer traveled by motor vehicles in the study area around the school in 2018
Kilometers traveled by motor vehicles in Barcelona in 2003-2017
Number of total trips and on foot trips, made by people in Barcelona 2003-2018
The other variables included were as follows:
Group: Intervention schools; Comparison schools
Period: pre-intervention; post-intervention
Year: 2002 to 2019


Statistical analysis
First, a descriptive analysis of the characteristics of the school’s surrounding area (neighborhood or study area around the school) for both the intervention and comparison schools, was carried out using the annual mean and its 95% confidence interval (95%CI) and the median and interquartile range. A descriptive analysis of the 6 dependent variables was also performed in each period (pre- and post-intervention) for both the intervention and comparison schools, for each school and for the total number of schools (aggregated), using the annual minimum, maximum, mean with its 95%CI and median with its interquartile range. The results between the intervention and comparison schools, both in the pre- and post-intervention periods, were compared using the non-parametric Wilcoxon rank-sum test (Mann-Whitney).
To compare the results in the post-intervention period versus the pre-period and to assess the effectiveness of the SRTS program in increasing road safety, for each outcome measure a GLMM with Poisson distribution using the logarithm as a link function between expected values and explanatory variables were fitted. The model can be formulated as:
log⁡(E(Y_i ))=X_i β+Z_i γ
Y_i∼Poi(μ_i)
where i is the observation index, Y is the response variable, X is the design matrix for fixed effects, is the vector of model coefficients, Z is the design matrix for the random effects and is a vector of normally distributed random coefficients.
The explanatory variables included in the model were the group, period, the interaction between group and period, and the year. To obtain a more precise fit, the model was also adjusted by the exposure variables. Among all the available exposure variables, those finally included in the model were: the number of students of the school (log), number of inhabitants in the neighborhood (log), kilometers traveled by motor vehicles in the study area (log), and a quadratic term for the latter effect (the relationship between the volume of traffic and the rate of accidents is not linear). These adjustment variables represented all the sources of variation we wanted to include and offered a better fit compared with: number of students, inhabitants in the neighborhood and kilometers traveled by motor vehicles.
To take into account the correlation between measurements and obtain a valid inference, various random factors were also evaluated: school, matched pair and year, and also random slopes for year at the school level.
For each outcome measure, the final model was selected based on the smallest Akaike Information Criteria, removing non-relevant random components.
Finally, in case the model showed overdispersion, the model was fitted again considering a GLMM with negative binomial distribution.
All models were validated by revising the absence of pattern in the residual plots against the predicted values.
Annual average estimates, relative risks (RR) and their 95% confidence intervals (95%CI) were obtained from the model least square means (LSMEANS, also called empirical marginal means), i.e. point estimates of different levels of interest evaluated as the average of other explanatory variables or random effects. The inverse transformation of the link function was applied to provide the results in the response scale.
To quantify the impact of the intervention, from the RRs, the number of traffic collisions, the number of collisions involving any injured 0-16 year-old, the number of collisions involving any injured 0-16 years-old pedestrians, the number of injured people, the number of 0-16 years-old injured and the number of 0-16 years-old pedestrians injured, in the post-intervention period versus the pre-intervention period were obtained, as the % change = -(1-RR)). Prevented cases were estimated from the difference in the % change between the intervention and the comparison schools, multiplied by the collisions in the first period of the intervention group.

Results
The study included 64 schools with SRTS programs implemented between 2006 and 2016, and 63 comparison schools, reaching 49,092 students in 2018. A total of 15.0% of schools were pre-schools (0-3 years), 55.1% were primary schools (4-11 years), and 29.9% were secondary schools (12-18 years).
In the intervention schools overall, the total number of people injured was 2,994 (annual mean of 272.2) in the pre-intervention period, and 2,284 (annual mean of 228.4) in the post-intervention period. In the comparison schools, this number was 4,061 (annual mean of 369.2) and 3,196 (annual mean of 319.6) respectively.
When we compared the results of the pre- and post-intervention periods, the final adjusted models showed a significant reduction in the risk of collisions and of injury in the intervention schools, with a percentage reduction of -11.7% in the number of collisions, -41.1% in the number of collisions involving 0-16 year-old. injured, and -43.3% in the number of collisions involving 0-16 years-old injured pedestrians. For people injured, these percentages were of -9.1% reductions in the total injured, -36.6% in that of 0–16-years-old injured, and -39.9% in the number of injured 0-16 years-old pedestrians.
Among the comparison schools, there were no significant changes in outcomes between the pre- and post-intervention periods.

The significant difference in percentage change in the post- versus the pre-period between intervention and comparison schools (significance of the interaction between intervention group and period) showed that the reduction in the intervention schools in the number of collisions injuring 0-16 years-old and 0-16 years-old pedestrians could be attributable to the implementation of the SRTS program.

Evaluation

What has been the effect of the activities?

The SRTS program, carried out in Barcelona between 2009 and 2016, showed a significant reduction in all groups in the intervention schools, which was not observed in the comparison schools. There was a greater decrease in the number of injured pedestrians, especially school-age pedestrians, which is the target population of the SRTS.

With the SRT program, each year it has prevented 16 collisions injuring 0-16 year-old, 15 collisions injuring 0-16 years-old pedestrians and 12 injured 0-16 years-old pedestrians. If the intervention had been expanded to all schools in the city, the number of prevented collisions and injured persons annually would have been 97, 90, and 75, respectively. Available data allowed to evaluate a real-life policy with important public health implications.

The results of our study show a reduction in the number of collisions and people injured in the SRTS intervention zones while the trend in the rest of the city was different. In Barcelona, from 2009 to 2019, the number of road traffic collisions involving injuries in the city increased by 5.3% (8,791 in 2009, 9,255 in 2019), and the number of people injured increased by 4.3% (11,461 in 2009, 11,838 in 2019). Road traffic injury rates per 100,000 vehicle-kilometer traveled increased during the same period by 7%.

This study shows that the SRTS program in Barcelona, has been effective in significantly reducing the number children injured around the school environment. Provides evidence for a policy that protects vulnerable road users as childhood. It provides support to develop SRTS not only in Barcelona, but also to many cities and contributes to improve road safety at urban level. This evaluation also supports the current extension of SRTS in Barcelona, under the new program “Protegim les escoles” (Let’s protect the schools).

Please briefly explain why your initiative is a good example of improving road safety.

These results are of relevance for two reasons. On the one hand, injuries were significantly reduced in the intervention schools but not in the comparison group, in a context of increasing road traffic injury rates in the city. On the other hand, because our results provided evidence of the effectiveness of the SRTS program in improving road safety and reducing road crashes and injuries, particularly among children, while there is controversy in the scientific literature. Our study aimed to overcome the limitations reported in previous studies by using a quasi-experimental study, which controlled for major confounding factors through the study design and statistical analysis.

How have you shared information about your project and its results?

We have shared the results at Municipal level with the departments of Mobility, Education, and Police. At national level we sent the results with the National Traffic Authority (Dirección General de Tráfico, DGT). At international level we share the results with Parnership for Healthy Cities, and International Transport Forum.

Disesemination done up until today includes:

Published papers:

Pérez K, Santamariña-Rubio E, Ferrando J, López MJ, Badiella L. Effectiveness of a Road Traffic Injury Prevention Intervention in Reducing Pedestrian Injuries, Barcelona, Spain, 2002-2019. Am J Public Health. 2023 May;113(5):495-499. doi: 10.2105/AJPH.2022.307216. Epub 2023 Feb 23.

Pérez K, Santamariña-Rubio E, Ferrando J, López MJ, Badiella L. Effectiveness of a Road Traffic Injury Prevention Intervention in Reducing Pedestrian Injuries, Barcelona, Spain, 2002-2019. Am J Public Health. 2023 May;113(5):495-499. doi: 10.2105/AJPH.2022.307216. Epub 2023 Feb 23.

Conference presentations:

Katherine Pérez, Elena Santamariña, Josep Ferrando, Maria José López Llorenç Badiella. Effectiveness of the Safe Routes to School program ("Camí
Escolar, Espai Amic") in reducing pedestrian injuries: a quasiexperimental
Study.

Scientific Meeting CIBERESP 2023. Parc Sanitari Sant Joan de Déu, Sant Boi de Llobregat. 27 y 28 de March de 2023. https://jornadacientifica.ciberesp.es/programa

2023 Presentation of the International Journal of Social Determinants of Health and Health Services (IJSDOHS). Universitat Pompeu Fabra, Barcelona 19 May 2023
Conference Posters: International Journal of Social Determinants of Health and Health Services: SAGE Journals (sagepub.com)

Abastract sent to:
EU Safety - 11th European Conference on Injury Prevention and Safety Promotion
5-6 October 2023 Reykjavik, Iceland


Press Release:

Municipal website
https://ajuntament.barcelona.cat/premsa/2023/03/30/els-camins-escolars-redueixen-un-40-el-nombre-de-vianants-infants-i-adolescents-atropellats-a-barcelona/

CIBER (Centro de Investigación Biomédica en Red, a Spanish public consortium, whose main purpose is the promotion and protection of health through the development of research in different areas.
https://www.ciberisciii.es/noticias/los-caminos-escolares-reducen-un-40-el-numero-de-peatones-ninos-y-adolescentes-atropellados-en-barcelona


Twitter:
https://twitter.com/salutpublicabcn/status/1641401314465726464
https://twitter.com/ciber_esp/status/1641377636650741760?s=48&t=8lTcdw3XRl772lmrzv6zmA
https://twitter.com/catherineprz1/status/1641469304246771714

In the news:
https://www.lavanguardia.com/sociedad/20230330/8864909/caminos-escolares-reducen-40-numero-peatones-ninos-adolescentes-atropellados-barcelona-estudio-agencia-salud-publica.html
https://www.elperiodico.com/es/barcelona/20230330/atropellos-menores-barcelona-entornos-escolares-bajan-85405037
https://www.larepublica.cat/minut-a-minut/els-camins-escolars-redueixen-un-40-els-atropellaments-a-infants-i-adolescents-a-barcelona/
https://www.servimedia.es/noticias/caminos-escolares-reducen-40-numero-peatones-ninos-adolescentes-atropellados-barcelona-estudio-agencia-salud-publica/3655428
https://cronicaglobal.elespanol.com/vida/se-reducen-atropellos-entornos-escolares-barcelona_788546_102.html