Multi-animal pose estimation, identification and tracking with DeepLabCut.

Estimating the pose of multiple animals is a challenging computer vision problem: frequent interactions cause occlusions and complicate the association of detected keypoints to the correct individuals, as well as having highly similar looking animals that interact more closely than in typical multi-human scenarios. To take up this challenge, we build on DeepLabCut, an open-source pose estimation toolbox, and provide high-performance animal assembly and tracking-features required for multi-animal scenarios. Furthermore, we integrate the ability to predict an animal's identity to assist tracking (in case of occlusions). We illustrate the power of this framework with four datasets varying in complexity, which we release to serve as a benchmark for future algorithm development.

Early Detection of Amyloid-Related Changes in Memory among Cognitively Unimpaired Older Adults with Daily Digital Testing.

OBJECTIVE: This study was undertaken to determine whether assessing learning over days reveals Alzheimer disease (AD) biomarker-related declines in memory consolidation that are otherwise undetectable with single time point assessments. METHODS: Thirty-six (21.9%) cognitively unimpaired older adults (aged 60-91 years) were classified with elevated ß-amyloid (Aß+) and 128 (78%) were Aß- using positron emission tomography with 11C Pittsburgh compound B. Participants completed the multiday Boston Remote Assessment for Neurocognitive Health (BRANCH) for 12 min/day on personal devices (ie, smartphones, laptops), which captures the trajectory of daily learning of the same content on 3 repeated tests (Digit Signs, Groceries-Prices, Face-Name). Learning is computed as a composite of accuracy across all 3 measures. Participants also completed standard in-clinic cognitive tests as part of the Preclinical Alzheimer's Cognitive Composite (PACC-5), with 123 participants undergoing PACC-5 follow-up after 1.07 (standard deviation = 0.25) years. RESULTS: At the cross-section, there were no statistically significant differences in performance between Aß+/- participants on any standard in-clinic cognitive tests (eg, PACC-5) or on day 1 of multiday BRANCH. Aß+ participants exhibited diminished 7-day learning curves on multiday BRANCH after 4 days of testing relative to Aß- participants (Cohen d = 0.49, 95% confidence interval = 0.10-0.87). Diminished learning curves were associated with greater annual PACC-5 decline (r = 0.54, p < 0.001). INTERPRETATION: Very early Aß-related memory declines can be revealed by assessing learning over days, suggesting that failures in memory consolidation predate other conventional amnestic deficits in AD. Repeated digital memory assessments, increasingly feasible and uniquely able to assess memory consolidation over short time periods, have the potential to be transformative for detecting the earliest cognitive changes in preclinical AD. ANN NEUROL 2024;95:507-517.

Development and validation of a questionnaire measuring knowledge, attitudes, and current practices of primary healthcare physicians regarding road injury prevention in older adults.

BACKGROUND: Older adults present high risk of involvement in road crashes. Preventive interventions conducted by their primary healthcare physicians (PHPC) could reduce this public health issue. OBJECTIVE: The objective of this study was to design and validate a self-administered questionnaire that measures the knowledge, attitudes, and current practices (CP) of PHCP in Spain regarding the prevention of road injuries in older adults. METHODS: One thousand eight hundred and ninety-seven PHCP completed a questionnaire piloted previously in an expert panel and two convenience samples of physicians. It comprised 78 items grouped in five sections and was mainly focused on exploring three constructs: knowledge, attitudes, and CP. Exploratory factor analysis was used to obtain evidence of internal structure validity. Reliability was assessed through Cronbach's α coefficient. Correlation coefficients for the scores constructed for each of the extracted factors were calculated to assess convergent and discriminant validity. RESULTS: Factor analysis extracted four factors each for the knowledge and attitudes constructs, and three factors for the CP construct, which explained more than 55% of the variance in each construct. Except for two factors of the knowledge construct regarding existing health problems associated to the risk of involvement in road crashes, the clustering pattern of all other items across the remaining nine factors was consistent and in agreement with previous knowledge. Cronbach's α values were greater than 0.7 for all constructs. CONCLUSIONS: Our questionnaire appears to be valid enough to assess the attitudes, CP, and medication-related knowledge of PHCP in Spain regarding the prevention of road injuries in older adults.

The clustering of physical activity and screen time behaviours in early childhood and impact on future health-related behaviours: a longitudinal analysis of children aged 3 to 8 years.

BACKGROUND: Meeting physical activity and screen time guidelines has been associated with improved health in children. Research has shown that lifestyle behaviours happen in combination and can be tracked into later life. Thus, a complex approach is needed to identify the effects of physical activity and screen time altogether. This study aims to identify clusters of both behaviours in a cohort of Irish 3-year-old children (n = 8833) and determine the association with sociodemographic characteristics and behaviours at age 5 and 7-8. METHODS: Data from the "Growing Up in Ireland" study collected between 2010 and 2016 was used in this study. Two-step cluster analysis was used to understand how physical activity and recreational screen time behaviours group together among 3-year-old children. Binary logistic regressions were conducted to examine if cluster placement at age 3 determined physical activity and recreational screen time behaviours at age 5 and 7-8 years, while controlling for gender of child, gender, age and employment status of the primary caregiver. RESULTS: Six clusters were identified in 9771 (49.3% female) 3-year-old children with the majority falling into a "High Active & Mixed Screen Time" (23.2%). Those in the "High Active & Mixed Screen Time" cluster at age 3 were more likely to engage in all physical activities reported at age 5 (p < 0.01) and age 7-8 (p < 0.01) when compared to a "Low Active & Screen Time Exceed" cluster. Children categorised in a "Moderate Active & Screen Time Below" and "Moderate Active & Screen Time Exceed" were more likely to engage in the same physical activities at age 5 and 7-8 (p < 0.05 - p < 0.01). However, children in the latter cluster were also more likely (p < 0.05) to play on a computer or tablet device. CONCLUSIONS: This paper highlights the importance of establishing positive health-related behaviours during early childhood, as this predicts future engagement in health-promoting activities. Regardless of screen time level, being part of a cluster with moderate or high levels of physical activity positively influences a child's future physical activity at age 5 and again at age 7 -8 years. The multiple layers of influence on a child's development should be leveraged to support the adoption of health-enhancing behaviours.

Different neighborhood walkability indexes for active commuting to school are necessary for urban and rural children and adolescents.

BACKGROUND: Literature focusing on youth has reported limited evidence and non-conclusive associations between neighborhood walkability measures and active commuting to and from school (ACS). Moreover, there is a lack of studies evaluating both macro- and micro-scale environmental factors of the neighborhood when ACS is analyzed. Likewise, most studies on built environment attributes and ACS focus on urban areas, whereas there is a lack of studies analyzing rural residential locations. Moreover, the relationship between built environment attributes and ACS may differ in children and adolescents. Hence, this study aimed to develop walkability indexes in relation to ACS for urban and rural children and adolescents, including both macro- and micro-scale school-neighborhood factors. METHODS: A cross-sectional study of 4593 participants from Spain with a mean age of 12.2 (SD 3.6) years was carried out. Macro-scale environmental factors were evaluated using geographic information system data, and micro-scale factors were measured using observational procedures. Socio-demographic characteristics and ACS were assessed with a questionnaire. Several linear regression models were conducted, including all the possible combinations of six or less built environment factors in order to find the best walkability index. RESULTS: Analyses showed that intersection density, number of four-way intersections, and residential density were positively related to ACS in urban participants, but negatively in rural participants. In rural children, positive streetscape characteristics, number of regulated crossings, traffic calming features, traffic lanes, and parking street buffers were also negatively related to ACS. In urban participants, other different factors were positively related to ACS: number of regulated crossings, positive streetscape characteristics, or crossing quality. Land use mix acted as a positive predictor only in urban adolescents. Distance to the school was a negative predictor on all the walkability indexes. However, aesthetic and social characteristics were not included in any of the indexes. CONCLUSIONS: Interventions focusing on improving built environments to increase ACS behavior need to have a better understanding of the walkability components that are specifically relevant to urban or rural samples.

Capturing learning curves with the multiday Boston Remote Assessment of Neurocognitive Health (BRANCH): Feasibility, reliability, and validity.

OBJECTIVE: Unsupervised remote digital cognitive assessment makes frequent testing feasible and allows for measurement of learning over repeated evaluations on participants' own devices. This provides the opportunity to derive individual multiday learning curve scores over short intervals. Here, we report feasibility, reliability, and validity, of a 7-day cognitive battery from the Boston Remote Assessment for Neurocognitive Health (Multiday BRANCH), an unsupervised web-based assessment. METHOD: Multiday BRANCH was administered remotely to 181 cognitively unimpaired older adults using their own electronic devices. For 7 consecutive days, participants completed three tests with associative memory components (Face-Name, Groceries-Prices, Digit Signs), using the same stimuli, to capture multiday learning curves for each test. We assessed the feasibility of capturing learning curves across the 7 days. Additionally, we examined the reliability and associations of learning curves with demographics, and traditional cognitive and subjective report measures. RESULTS: Multiday BRANCH was feasible with 96% of participants completing all study assessments; there were no differences dependent on type of device used (t = 0.71, p = .48) or time of day completed (t = -0.08, p = .94). Psychometric properties of the learning curves were sound including good test-retest reliability of individuals' curves (intraclass correlation = 0.94). Learning curves were positively correlated with in-person cognitive tests and subjective report of cognitive complaints. CONCLUSIONS: Multiday BRANCH is a feasible, reliable, and valid cognitive measure that may be useful for identifying subtle changes in learning and memory processes in older adults. In the future, we will determine whether Multiday BRANCH is predictive of the presence of preclinical Alzheimer's disease. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Cycling area can be a confounder and effect modifier of the association between helmet use and cyclists' risk of death after a crash.

The effect of helmet use on reducing the risk of death in cyclists appears to be distorted by some variables (potential confounders, effect modifiers, or both). Our aim was to provide evidence for or against the hypothesis that cycling area may act as a confounder and effect modifier of the association between helmet use and risk of death of cyclists involved in road crashes. Data were analysed for 24,605 cyclists involved in road crashes in Spain. A multiple imputation procedure was used to mitigate the effect of missing values. We used multilevel Poisson regression with province as the group level to estimate the crude association between helmet use and risk of death, and also three adjusted analyses: (1) for cycling area only, (2) for the remaining variables which may act as confounders, and (3) for all variables. Incidence-density ratios (IDR) and their 95% confidence intervals were calculated. Crude IDR was 1.10, but stratifying by cycling area disclosed a protective, differential effect of helmet use: IDR = 0.67 in urban areas, IDR = 0.34 on open roads. Adjusting for all variables except cycling area yielded similar results in both strata, albeit with a smaller difference between them. Adjusting for cycling area only yielded a strong association (IDR = 0.42), which was slightly lower in the adjusted analysis for all variables (IDR = 0.45). Cycling area can act as a confounder and also appears to act as an effect modifier (albeit to a lesser extent) of the risk of cyclists' death after a crash.

The Effect of a School-Based Intervention on Children's Cycling Knowledge, Mode of Commuting and Perceived Barriers: A Randomized Controlled Trial.

The low rates of active commuting to/from school in Spain, especially by bike, and the wide range of cycling interventions in the literature show that this is a necessary research subject. The aims of this study were: (1) to assess the feasibility of a school-based cycling intervention program for adolescents, (2) to analyse the effectiveness of a school-based cycling intervention program on the rates of cycling and other forms of active commuting to/from school (ACS), and perceived barriers to active commuting in adolescents. A total of 122 adolescents from Granada, Jaén and Valencia (Spain) participated in the study. The cycling intervention group participated in a school-based intervention program to promote cycling to school during Physical Education (PE) sessions in order to analyse the changes in the dependent variables at baseline and follow up of the intervention. Wilcoxon, Signs and McNemar tests were undertaken. The association of the intervention program with commuting behaviour, and perceived barriers to commuting, were analysed by binary logistic regression. There were improvements in knowledge at follow-up and the cycling skill scores were medium-low. The rates of cycling to school and active commuting to/from school did not change, and only the "built environment (walk)" barrier increased in the cycling group at follow-up. School-based interventions may be feasibly effective tools to increase ACS behaviour, but it is necessary to implement a longer period and continue testing further school-based cycling interventions.

Convergent Validation of a Self-Reported Commuting to and from School Diary in Spanish Adolescents.

The aim of this study was to examine the convergent validity of self-reported diary times for commuting to and from school with device-measured positional data (Global Positioning System; GPS) in Spanish adolescents. METHODS: Cross-sectional data were obtained from four Spanish public secondary schools in 2021, comprising 47 adolescents and 141 home-school and school-home trips. Participants self-reported the time they left and arrived at home and school through a commuting diary. They wore a GPS device recording the objective time during three trips (i.e., one home-school trip and two school-home trips). Agreement between commuting diary and GPS data regarding home-school trips and school-home trips was evaluated using Bland-Altman plots. RESULTS: Total commuting time differed by 1 min (95% limits of agreement were 16.1 min and -18.1 min) between subjective and objective measures (adolescents reported 0.8 more minutes in home-school trips and 1 more minute in school-home trips compared to objective data). Passive commuters reported 0.7 more minutes and active commuters reported 1.2 more minutes in the total commuting time compared to objective data. CONCLUSIONS: Self-reported commuting diaries may be a useful tool to obtain commuting times of adolescents in epidemiological research or when tools to measure objective times are not feasible.

How many collateral casualties are associated with road users responsible for road crashes?

OBJECTIVE: To estimate the number of collateral casualties associated with road users considered responsible for a road crash. METHOD: We analyzed the case series comprising all 790,435 road users involved in road crashes with victims in Spain from 2009 to 2013, recorded in a nationwide police-based registry. For each road user assumed to be responsible for a crash, we collected information relative to health outcomes in other people involved in it, and obtained the total number of collateral casualties per 100 road users considered responsible for the crash. We then estimated the strength of associations between sex, age and the number of collateral casualties generated by car drivers considered responsible for the crash, and calculated rate ratios and corresponding 95% confidence intervals. RESULTS: Pedestrians responsible for crashes were associated with the lowest number of collateral casualties (13.1/100), whereas the highest number (153/100) was observed for bus drivers responsible for crashes. Car drivers were associated with 104.4/100 collateral casualties. The youngest and the oldest car drivers responsible for crashes were associated with 33% and 41% more deaths, respectively, than the 25-34 year old group. Male drivers were associated with 22% more collateral casualties than female drivers. CONCLUSIONS: Regardless of the type of road user who was responsible for a road crash, their active contribution to the crash led to an additional number of collateral casualties in other, non-responsible users. The number and severity of collateral casualties were related to the type of vehicle and the number of people involved. These results are potentially useful to support the need to promote safer driver practices among subgroups of high-risk drivers.

Unsupervised mobile cognitive testing for use in preclinical Alzheimer's disease.

INTRODUCTION: Unsupervised digital cognitive testing is an appealing means to capture subtle cognitive decline in preclinical Alzheimer's disease (AD). Here, we describe development, feasibility, and validity of the Boston Remote Assessment for Neurocognitive Health (BRANCH) against in-person cognitive testing and amyloid/tau burden. METHODS: BRANCH is web-based, self-guided, and assesses memory processes vulnerable in AD. Clinically normal participants (n = 234; aged 50-89) completed BRANCH; a subset underwent in-person cognitive testing and positron emission tomography imaging. Mean accuracy across BRANCH tests (Categories, Face-Name-Occupation, Groceries, Signs) was calculated. RESULTS: BRANCH was feasible to complete on participants' own devices (primarily smartphones). Technical difficulties and invalid/unusable data were infrequent. BRANCH psychometric properties were sound, including good retest reliability. BRANCH was correlated with in-person cognitive testing (r = 0.617, P < .001). Lower BRANCH score was associated with greater amyloid (r = -0.205, P = .007) and entorhinal tau (r = -0.178, P = .026). DISCUSSION: BRANCH reliably captures meaningful cognitive information remotely, suggesting promise as a digital cognitive marker sensitive early in the AD trajectory.

A School-Based Randomized Controlled Trial to Promote Cycling to School in Adolescents: The PACO Study.

This manuscript describes the rationale and protocol of a school-based randomized controlled trial called "Cycling and Walk to School" (PACO, by its Spanish acronym) that aims to promote cycling to and from school and physical activity (PA) in adolescents. This study will examine the effects of this intervention in cycling and active commuting to and from school (ACS), PA and several ACS-related factors based on self-determination theory (SDT) and a social-ecological model (SEM). A total of 360 adolescents attending six high schools (three experimental and three control) from three Spanish cities will participate in this randomized controlled trial. The intervention (four cycling sessions; 1-2 h per session, one session per week) will be conducted by the research staff; the control group will continue their usual activities. PA levels will be measured by accelerometers, whereas ACS and the other study variables will be self-reported using questionnaires at baseline and post-intervention. The primary outcomes will be: rates of cycling to school, ACS and PA levels. In addition, SDT-related variables and individual, interpersonal, community, and environment variables relevant to ACS will be based on SEM. The findings will provide a comprehensive understanding of the short-term effects of this school-based intervention on cycling to school behaviour, ACS and PA levels in Spanish adolescents.

[Association between type of vehicle and the risk of provoking a collision between vehicles]. / Asociación del tipo de vehículo con el riesgo de provocar una colisión entre vehículos.

OBJECTIVE: To quantify the magnitude of the association between the type of vehicle and the probability of being responsible for a collision between two or more vehicles. METHOD: From the registry of road crashes with victims maintained by the Spanish Traffic General Directorate (2014 and 2015), a matched case-control study was designed. Cases were offending drivers involved in the 27,630 collisions between two or more vehicles in which only one of the drivers had committed a driving mistake or offence. Each case was matched with the non-offending drivers of the vehicles involved in the same crash: in all, 31,219 controls were included. Apart from the commission of offences and the type of vehicle involved, we got information about other characteristics of the driver (age, sex, etc.) and about the vehicle (age). Odds ratios (OR) were calculated in order to quantify the association between each type of vehicle and the odds of being responsible for the collision, crude and adjusted (by conditioned logistic regression) by the rest of collected variables. RESULTS: In comparison with private cars, bicycles had a lower risk of causing a collision (adjusted OR: .30), and also mopeds (aOR: .52) and buses (aOR: .63). Vans (aOR: 1.19) and four-wheel vehicles (aOR: 1.33) increased the risk. CONCLUSION: Two-wheeled vehicles and buses had a lower risk of causing collisions than private cars. This association is independent of some of the characteristics of the driver, as well as the age of the vehicle.

Individual and environmental factors associated with death of cyclists involved in road crashes in Spain: a cohort study.

OBJECTIVE: To quantify the magnitude of associations between cyclist fatalities and both cyclist and environment related characteristics in Spain during the first 24 hours after a crash. DESIGN: Cohort study. SETTING: Spain. PARTICIPANTS: 65 977 cyclists injured in road crashes recorded between 1993 and 2013 in the Spanish Register of Road Crashes with Victims. MAIN OUTCOME: Death within the first 24 hours after the crash. METHODS: A multiple imputation procedure was used to mitigate the effect of missing values. Differences between regions were assumed and managed with multilevel analysis at the cyclist and province levels. Incidence density ratios (IDR) with 95% CI were calculated with a multivariate Poisson model. RESULTS: Non-use of a helmet was directly associated with death (IDR 1.43, 95% CI 1.25 to 1.64). Among other cyclist characteristics, age after the third decade of life was also directly associated with death, especially in older cyclists ('over 74' category, IDR 4.61, 95% CI 3.49 to 6.08). The association with death did not differ between work-related cycling and other reasons for cycling.There was an inverse association with death for crashes in urban areas and on community roads. Any adverse meteorological condition also showed a direct association with death, whereas altered road surfaces showed an inverse association. Crashes during nighttime were directly associated with death, with a peak between 3:00 and 5:59 am (IDR 1.58, 95% CI 1.03 to 2.41). CONCLUSIONS: We found strong direct and inverse associations between several cyclist and environment related variables and death. These variables should be considered in efforts to prioritise public health measures aimed at reducing the number of cycling-related fatalities.

How many collateral casualties are associated with road users responsible for road crashes? / ¿Cuántas víctimas colaterales se asocian con los usuarios de la carretera responsables de los accidentes de tráfico?

Objective: To estimate the number of collateral casualties associated with road users considered responsible for a road crash. Method: We analyzed the case series comprising all 790,435 road users involved in road crashes with victims in Spain from 2009 to 2013, recorded in a nationwide police-based registry. For each road user assumed to be responsible for a crash, we collected information relative to health outcomes in other people involved in it, and obtained the total number of collateral casualties per 100 road users considered responsible for the crash. We then estimated the strength of associations between sex, age and the number of collateral casualties generated by car drivers considered responsible for the crash, and calculated rate ratios and corresponding 95% confidence intervals. Results: Pedestrians responsible for crashes were associated with the lowest number of collateral casualties (13.1/100), whereas the highest number (153/100) was observed for bus drivers responsible for crashes. Car drivers were associated with 104.4/100 collateral casualties. The youngest and the oldest car drivers responsible for crashes were associated with 33% and 41% more deaths, respectively, than the 25-34 year old group. Male drivers were associated with 22% more collateral casualties than female drivers. Conclusions: Regardless of the type of road user who was responsible for a road crash, their active contribution to the crash led to an additional number of collateral casualties in other, non-responsible users. The number and severity of collateral casualties were related to the type of vehicle and the number of people involved. These results are potentially useful to support the need to promote safer driver practices among subgroups of high-risk drivers. (AU)

Objetivo: Estimar el número de víctimas colaterales asociadas a los usuarios de las vías de tráfico responsables de una colisión. Método: Se estudió la serie de casos formada por los 790.435 usuarios de la vía implicados en accidentes de tráfico con víctimas en España, entre 2009 y 2013, recogidos en el Registro de Accidentes de Tráfico con Víctimas de la Dirección General de Tráfico. Para cada uno de los usuarios considerado responsable se recogió información sobre las consecuencias sanitarias para las otras personas implicadas en el mismo accidente, y se calculó el número total de víctimas colaterales por cada 100 usuarios responsables. A continuación se estimó la fuerza de asociación de la edad y el sexo de los conductores de turismo con el número de víctimas colaterales generadas por ellos, y se obtuvieron las correspondientes razones de tasas y sus intervalos de confianza del 95%. Resultados: Los peatones responsables de atropellos se asociaron al menor número de víctimas colaterales (13,1/100), mientras que la cifra más alta de estas (153/100) se asoció a los conductores de autobús responsables del accidente. Los de turismo implicaron a 104,4/100. Los conductores de turismo más jóvenes (<25 años) y los de mayor edad (>64 años) implicaron respectivamente un 33% y un 41% más de víctimas colaterales que el grupo de 25-34 años. Los varones se asociaron a un 22% más de víctimas colaterales que las mujeres. Conclusiones: Independientemente del tipo de usuario responsable del accidente, su contribución activa se asoció a un número adicional de víctimas colaterales, cuyas magnitud y gravedad dependieron sobre todo del tipo de vehículo conducido. Este trabajo puede ser útil para reforzar las actuaciones destinadas a prevenir la accidentalidad en los usuarios de alto riesgo. (AU)

Asociación del tipo de vehículo con el riesgo de provocar una colisión entre vehículos / Association between type of vehicle and the risk of provoking a collision between vehicles

OBJETIVO: Cuantificar la magnitud de la asociación entre el tipo de vehículo y la probabilidad de ser el responsable de una colisión entre dos o más vehículos. MÉTODO: A partir del registro de accidentes de tráfico con víctimas de la Dirección General de Tráfico (2014 y 2015) se diseñó un estudio de casos y controles emparejado. Los casos fueron los conductores infractores implicados en las 27.630 colisiones entre dos o más vehículos, en las que solo uno de los conductores implicados había cometido algún error de conducción o infracción. Cada caso se emparejó con los conductores no infractores de los vehículos implicados en el mismo accidente; en total se dispuso de 31.219 controles. Aparte de la comisión de infracciones y del tipo de vehículo implicado, se obtuvo información para otras características del conductor (edad, sexo, etc.) y del vehículo (antigüedad). Se calcularon odds ratios (OR) para cuantificar la asociación entre cada tipo de vehículo y la odds de ser el causante de la colisión, crudas y ajustadas (ORa) (regresión logística condicionada) por el resto de las variables recogidas. RESULTADOS: En comparación con los turismos, se obtuvo un menor riesgo de provocar la colisión para bicicletas (ORa: 0,30), ciclomotores (ORa: 0,52) y autobuses (ORa: 0,63), y un mayor riesgo para furgonetas (ORa: 1,19) y vehículos todoterreno (ORa: 1,33). CONCLUSIÓN: Los vehículos de dos ruedas y los autobuses tienen un menor riesgo de provocar colisiones que los turismos. Esta asociación es independiente de algunas características del conductor, así como de la antigüedad del vehículo

OBJECTIVE: To quantify the magnitude of the association between the type of vehicle and the probability of being responsible for a collision between two or more vehicles. METHOD: From the registry of road crashes with victims maintained by the Spanish Traffic General Directorate (2014 and 2015), a matched case-control study was designed. Cases were offending drivers involved in the 27,630 collisions between two or more vehicles in which only one of the drivers had committed a driving mistake or offence. Each case was matched with the non-offending drivers of the vehicles involved in the same crash: in all, 31,219 controls were included. Apart from the commission of offences and the type of vehicle involved, we got information about other characteristics of the driver (age, sex, etc.) and about the vehicle (age). Odds ratios (OR) were calculated in order to quantify the association between each type of vehicle and the odds of being responsible for the collision, crude and adjusted (by conditioned logistic regression) by the rest of collected variables. RESULTS: In comparison with private cars, bicycles had a lower risk of causing a collision (adjusted OR: .30), and also mopeds (aOR: .52) and buses (aOR: .63). Vans (aOR: 1.19) and four-wheel vehicles (aOR: 1.33) increased the risk. CONCLUSION: Two-wheeled vehicles and buses had a lower risk of causing collisions than private cars. This association is independent of some of the characteristics of the driver, as well as the age of the vehicle