Fatality risk for belted drivers versus car mass

This study was performed to determine how the likelihood of a belted driver being killed in a single car crash depends on the mass of the car. This was done by applying the pedestrian fatality exposure approach to the subset of fatalities in the Fatal Accident Reporting System (FARS) for which the driver was coded as using a shoulder belt and/or a lap belt. Combining the 1975 through 1982 data provided a sufficiently large population of belted drivers to perform the analysis. In the exposure approach used, the number of car drivers killed in single car crashes is divided by the number of nonoccupant fatalities (pedestrians or motorcyclists) associated with the same group of cars. The ratio is interpreted to reflect the physical effect of car mass, essentially independent of driver behavior effects. In the present application, car mass effects for belted drivers were determined by considering the number of belted drivers killed divided by the number of nonoccupants killed in crashes involving cars whose drivers were coded in the FARS files as being belted. Because the belt use of surviving drivers is, to some extent, self-reported, it is considered that the data given in the report should be not used to estimate the effectiveness of seat belts in preventing fatalities. The results are presented as graphical and analytical comparisons of fatality likelihood versus car mass for belted and unbelted drivers. It is concluded that the effect of car mass on relative driver fatality likelihood is essentially the same for belted and unbelted drivers (for example, the present analysis gives that a belted driver in a 900 kg car is 2.3 times as likely to be killed in a single car crash as is the belted driver in an 1800 kg car. The corresponding ratio determined here for unbelted drivers is 2.4). As a consequence of this conclusion, the relative effectiveness of seat belts in preventing driver fatalities is similar for cars of different masses.     

Serious or fatal driver injury rate versus car mass in head-on crashes between cars of similar mass

This work was performed to determine relations between car mass and driver injuries (serious or fatal) when cars of similar mass crash into each other head-on. This type of crash is examined because it is considered similar in some respects to a barrier crash. Data from the United States Fatal Accident Reporting System (FARS) are used to examine driver fatality likelihood as a function of car mass when cars of similar mass crash into each other. Pedestrian fatalities involving cars of the same mass are used to estimate exposure. Two additional sources of data (State data from North Carolina and New York) are used to generate information on the number of drivers seriously injured or killed per police reported crash when cars of similar mass crash into each other. The present study finds that the likelihood of driver injury (fatal or serious) when cars of similar mass crash into each other increases with decreasing car mass, both for head-on crashes and for crashes in all directions. The study does not address possible mechanisms that might lead to such relations. All the data analyzed reveal a fairly consistent picture--a driver in a 900 kg car crashing head-on into another 900 kg car is about 2.0 times as likely to be seriously injured or killed as is a driver of a 1800 kg car crashing head-on into another 1800 kg car.     

A note on the variance of paired comparisons estimates

The method of paired comparisons to estimate treatment effectiveness was introduced by Evans (Evans L. Double pair comparison--a new method to determine how occupant characteristics affect fatality risk in traffic crashes. Accident Analysis and Prevention 1985;12:217-27). It is similar in form to other effectiveness estimates based on odds ratios using independent groups. Therefore, it has been generally assumed that the variance is computed in the same way. In this note, it is demonstrated. using a simple binomial model and linear approximation, that the variance is lower for paired comparisons estimates than for odds ratios estimates based on independent groups. In order to use odds ratios, there must be a treated group and an untreated group. Within each group there are occurrences of an event against which treatment effectiveness is being estimated and, also, occurrences of different event that is considered unamenable to the treatment. The treatment effectiveness, e, is estimated by 1-R where R is the ratio of amenable type events to unamenable ones in the treated group, divided by the same ratio for the untreated group. A distinction is made between 'real' paired comparisons and odds ratios based on independent data. An example of the independent case is. x is the number of fatalities in frontal crashes without air bags; y the number of fatalities in non-frontal crashes without air bags; s the number of fatalities in frontal crashes with air bags, and t the number of fatalities in non-frontal crashes with air bags. While fatalities in non-frontal crashes serve as denominators in R, a particular frontal crash is not paired with one particular non-frontal crash. In this case, in which all the data are independent, the variance of e is approximately R2(1/x + 1/y + 1/s + 1/t), a result which is consistent with well known results about the log odds ratio. For an example of real paired comparisons, we consider fatalities in cars that have a driver and exactly one unbelted right front seat passenger. Suppose there were x driver fatalities and y passenger fatalities in the cars in which the driver was also unbelted and s driver fatalities and t passenger fatalities in the cars in which the driver was belted. Since the fate of the passenger would not be amenable to 'treating' the driver, the same estimate of belt effectiveness based on these data. 1 - (s/t)/(x/y), is reasonable. In this case, x and y, and s and t are not independent. This is due to the fact that while the overall probability of fatality in a crash is very low, the conditional probability of fatality given that someone else in the car died is greater than the unconditional probability of fatality. Under these circumstances, the variance of the paired comparisons estimate is reduced. Published by Elsevier Science Ltd.     

Crash involvement of drivers with multiple crashes

A goal for any licensing agency is the ability to identify high-risk drivers. Kentucky data show that a significant number of drivers are repeatedly involved in crashes. The objective of this study is the development of a crash prediction model that can be used to estimate the likelihood of a driver being at fault for a near future crash occurrence. Multiple logistic regression techniques were employed using the available data for the Kentucky licensed drivers. This study considers as crash predictors the driver's total number of previous crashes, citations accumulated, the time gap between the latest two crashes, crash type, and demographic factors. The driver's total number of previous crashes was further disaggregated into the drivers' total number of previous at-fault and not-at-fault crashes. The model can be used to correctly classify at-fault drivers up to 74.56% with an overall efficiency of 63.34%. The total number of previous at-fault crash involvements, and having previous driver license suspensions and traffic school referrals are strongly associated with a driver being responsible for a subsequent crash. In addition, a driver's likelihood to be at fault in a crash is higher for very young or very old, males, drivers with both speeding and non-speeding citations, and drivers that had a recent crash involvement. Thus, the model presented here enables agencies to more actively monitor the likelihood of a driver to be at fault in a crash.     

Relative risk of injury and death in ambulances and other emergency vehicles

This study addresses of the impacts of emergency vehicle (ambulances, police cars and fire trucks) occupant seating position, restraint use and vehicle response status on injuries and fatalities. Multi-way frequency and ordinal logistic regression analyses were performed on two large national databases, the National Highway Traffic Safety Administration’s Fatality Analysis Reporting System (FARS) and the General Estimates System (GES). One model estimated the relative risk ratios for different levels of injury severity to occupants traveling in ambulances. Restrained ambulance occupants involved in a crash were significantly less likely to be killed or seriously injured than unrestrained occupants. Ambulance rear occupants were significantly more likely to be killed than front-seat occupants. Ambulance occupants traveling non-emergency were more likely than occupants traveling emergency to be killed or severely injured. Unrestrained ambulance occupants, occupants riding in the patient compartment and especially unrestrained occupants riding in the patient compartment were at substantially increased risk of injury and death when involved in a crash. A second model incorporated police cars and fire trucks. In the combined ambulance–fire truck–police car model, the likelihood of an occupant fatality for those involved in a crash was higher for routine responses. Relative to police cars and fire trucks, ambulances experienced the highest percentage of fatal crashes where occupants are killed and the highest percentage of crashes where occupants are injured. Lack of restraint use and/or responding with ‘lights and siren’ characterized the vast majority of fatalities among fire truck occupants. A third model incorporated non-special use van and passenger car occupants, which otherwise replicated the second model. Our findings suggest that ambulance crewmembers riding in the back and firefighters in any seating position, should be restrained whenever feasible. Family members accompanying ambulance patients should ride in the front-seat of the ambulance.     

Potential fatality reductions through eliminating occupant ejection from cars

The percent of occupant fatalities preventable by eliminating ejection is calculated using Fatal Accident Reporting System (FARS) data for 1975 through 1986. The calculation requires estimates of two quantities. First, the fraction of all fatally injured occupants who were ejected; this is obtained directly from the FARS data. Second, the probability that an ejected occupant was killed compared to the probability that the occupant would have been killed in a similar crash in the absence of ejection; this quantity is estimated using the double pair comparison method, and its dependence on occupant age and sex and on car mass and model year is examined. High precision estimates of the reduction in fatalities that would result from eliminating ejection as functions of these same variables are thereby obtained. These estimates depend on assuming that whatever method is used to prevent ejection would cause the formerly ejected occupant to acquire the same fatality risk as a nonejected occupant in a similar crash; the study does not address how to prevent ejection. It is concluded that ejection elimination would decrease fatalities to unrestrained car occupants by 18 +/- 1%. The fatality reductions are independent of car seating position (19%, 19%, 17%, 16%, 19%, and 18% for drivers, middle front, right front, left rear, middle rear, and right rear passengers, respectively); they decrease with driver age, from 25% at age 18 years to 7% at 70 years; they decrease with increasing mass, but remained relatively independent of car model year since the early 1970s, being somewhat higher for earlier model years.     

Increased risk of driver fatality due to unrestrained rear-seat passengers in severe frontal crashes

While belt usage among rear-seat passengers is disproportionately lower than their front-seat counterpart, this may have serious consequences in the event of a crash not only for the unbelted rear-seat passenger but also for the front-seat passengers as well. To quantify that effect, the objective of the study is to evaluate the increased likelihood of driver fatality in the presence of unrestrained rear-seat passengers in a severe frontal collision. U.S.-based census data from 2001 to 2009 fatal motor vehicle crashes was used to enroll frontal crashes which involved 1998 or later year vehicle models with belted drivers and at least one adult passenger in the rear left seat behind the driver. Results using multivariate logistic regression analysis indicated that the odds of a belt restrained driver sustaining a fatal injury was 137% (95% CI = 95%, 189%) higher when the passenger behind the driver was unbelted in comparison to a belted case while the effects of driver age, sex, speed limit, vehicle body type, airbag deployment and driver ejection were controlled in the model. The likelihood of driver fatality due to an unrestrained rear left passenger increased further (119–197%) in the presence of additional unrestrained rear seat passengers in the rear middle or right seats. The results from the study highlight the fact that future advances to front row passive safety systems (e.g. multi-stage airbag deployment) must be adapted to take into account the effect of unrestrained rear-seat passengers.     

Declines in fatal crashes of older drivers: changes in crash risk and survivability

Objectives

Previous research has found that older driver fatal crash involvement rates per licensed driver declined substantially in the United States during 1997–2006 and declined much faster than the rate for middle-age drivers. The current study examined whether the larger-than-expected decline for older drivers extended to nonfatal crashes and whether the decline in fatal crash risk reflects lower likelihood of crashing or an improvement in survivability of the crashes that occur.

Methods

Trends in the rates of passenger vehicle crash involvements per 100,000 licensed drivers for drivers 70 and older (older drivers) were compared with trends for drivers ages 35–54 (middle-age drivers). Fatal crash information was obtained from the Fatality Analysis Reporting System for years 1997–2008, and nonfatal crash information was obtained from 13 states with good reporting information for years 1997–2005. Analysis of covariance models compared trends in annual crash rates for older drivers relative to rates for middle-age drivers. Differences in crash survivability were measured in terms of the odds of fatality given a crash each year, and the historical trends for older versus middle-age drivers were compared.

Results

Fatal crash involvement rates declined for older and middle-age drivers during 1997–2008 (1997–2005 for the 13 state subsample), but the decline for drivers 70 and older far exceeded the decline for drivers ages 35–54 (37 versus 23 percent, nationally; 22 versus 1 percent, 13 states). Nonfatal injury crash involvement rates showed similarly larger-than-expected declines for older drivers in the 13 state subsample, but the differences were smaller and not statistically significant (27 percent reduction for older drivers versus 16 percent for middle-age drivers). Property-damage-only crash involvement rates declined for older drivers (10 percent) but increased for middle-age drivers (1 percent). In 1997, older drivers were 3.5 times more likely than middle-age drivers to die in police-reported crashes (6.2 versus 1.8 deaths per 1000 crashes), but this difference was reduced during the 9-year study period to 2.9 times, as the rate of older drivers dying in a crash declined (5.5 deaths per 1000 crashes in 2005) and the death risk remained relatively stable for middle-age drivers.

Conclusions

Contrary to expectations based on increased licensure and travel by older drivers, their fatal crash risk has declined during the past decade and has declined at a faster rate than for middle-age drivers. The decreased risk for older drivers appears to extend not only to nonfatal injury crashes but also to property-damage-only crashes, at least as reported to police in the 13 states included in the nonfatal injury analysis. Although insurance collision data suggest that overall crash risk of older drivers may not be changing relative to middle-age drivers, the current analysis indicates that the reduced fatality risk of older drivers reflects both less likelihood of being involved in a police-reported crash and greater likelihood that they will survive when they do crash.     

Age and pedestrian injury severity in motor-vehicle crashes: a heteroskedastic logit analysis

This research explores the injury severity of pedestrians in motor-vehicle crashes. It is hypothesized that the variance of unobserved pedestrian characteristics increases with age. In response, a heteroskedastic generalized extreme value model is used. The analysis links explanatory factors with four injury outcomes: fatal, incapacitating, non-incapacitating, and possible or no injury. Police-reported crash data between 1997 and 2000 from North Carolina, USA, are used. The results show that pedestrian age induces heteroskedasticity which affects the probability of fatal injury. The effect grows more pronounced with increasing age past 65. The heteroskedastic model provides a better fit than the multinomial logit model. Notable factors increasing the probability of fatal pedestrian injury: increasing pedestrian age, male driver, intoxicated driver (2.7 times greater probability of fatality), traffic sign, commercial area, darkness with or without streetlights (2-4 times greater probability of fatality), sport-utility vehicle, truck, freeway, two-way divided roadway, speeding-involved, off roadway, motorist turning or backing, both driver and pedestrian at fault, and pedestrian only at fault. Conversely, the probability of a fatal injury decreased: with increasing driver age, during the PM traffic peak, with traffic signal control, in inclement weather, on a curved roadway, at a crosswalk, and when walking along roadway.     

Limits of the quasi-induced exposure method when compared with the standard case-control design. Application to the estimation of risks associated with driving under the influence of cannabis or alcohol

The disparities between the quasi-induced exposure (QIE) method and a standard case–control approach with crash responsibility as disease of interest are studied. The 10,748 drivers who had been given compulsory cannabis and alcohol tests subsequent to involvement in a fatal crash in France between 2001 and 2003 were used to compare the two approaches. Odds ratios were assessed using conditional and unconditional logistic regressions. While both approaches found that drivers under the influence of alcohol or cannabis increased the risk of causing a fatal crash, the two approaches are not equivalent. They differ mainly with regards to the driver sample selected. The QIE method results in splitting the overall road safety issue into two sub-studies: a matched case–control study dealing with two-vehicle crashes and a case–control study dealing with single-vehicle crashes but with a specific control group. Using a specific generic term such as “QIE method” should not hide the real underlying epidemiological design. On the contrary, the standard case–control approach studies drivers involved in all type of crashes whatever the distribution of the responsibility in each crash. This method also known as “responsibility analysis” is the most relevant for assessing the overall road safety implications of a driver characteristic.     

Exploring the impacts of factors contributing to tram-involved serious injury crashes on Melbourne tram routes

Previous research is limited regarding factors influencing tram-involved serious injury crashes. The aim of this study is to identify key vehicle, road, environment and driver related factors associated with tram-involved serious injury crashes. Using a binary logistic regression modelling approach, the following factors were identified to be significant in influencing tram-involved fatal crashes in Melbourne: tram floor height, tram age, season, traffic volume, tram lane priority and tram travel speed. Low floor trams, older trams, tram priority lanes and higher tram travelling speeds are more likely to increase tram-involved fatal crashes. Higher traffic volume decreases the likelihood of serious crashes. Fatal crashes are more likely to occur during spring and summer. Findings from this study may offer ideas for future research in the area of tram safety and help to develop countermeasures to prevent specific fatality types from occurring.     

Using speeding detections and numbers of fatalities to estimate relative risk of a fatality for motorcyclists and car drivers

Precise estimation of the relative risk of motorcyclists being involved in a fatal accident compared to car drivers is difficult. Simple estimates based on the proportions of licenced drivers or riders that are killed in a fatal accident are biased as they do not take into account the exposure to risk. However, exposure is difficult to quantify. Here we adapt the ideas behind the well known induced exposure methods and use available summary data on speeding detections and fatalities for motorcycle riders and car drivers to estimate the relative risk of a fatality for motorcyclists compared to car drivers under mild assumptions. The method is applied to data on motorcycle riders and car drivers in Victoria, Australia in 2010 and a small simulation study is conducted.     

Rollover risk of cars and light trucks after accounting for driver and environmental factors

Characteristics of the driver, roadway environment, and vehicle were associated with the likelihood of rollover occurrence in more than 14000 single-vehicle fatal and 78000 single-vehicle injury crashes during 1995-98. Rollovers were more likely in crashes involving young drivers or occurring on rural curves. After accounting for the effects of driver age and gender, roadway alignment and surface condition, and whether or not the crash occurred in a rural area, light trucks were still twice as likely as cars to experience rollovers. Some light truck models were much more likely than others to experience rollovers. However, while physical differences (e.g. center of gravity height) could explain some of this variability, other factors affecting vehicle stability may be evident only after dynamic testing.     

Driver injury severity outcome analysis in rural interstate highway crashes: a two-level Bayesian logistic regression interpretation

There is a high potential of severe injury outcomes in traffic crashes on rural interstate highways due to the significant amount of high speed traffic on these corridors. Hierarchical Bayesian models are capable of incorporating between-crash variance and within-crash correlations into traffic crash data analysis and are increasingly utilized in traffic crash severity analysis. This paper applies a hierarchical Bayesian logistic model to examine the significant factors at crash and vehicle/driver levels and their heterogeneous impacts on driver injury severity in rural interstate highway crashes. Analysis results indicate that the majority of the total variance is induced by the between-crash variance, showing the appropriateness of the utilized hierarchical modeling approach. Three crash-level variables and six vehicle/driver-level variables are found significant in predicting driver injury severities: road curve, maximum vehicle damage in a crash, number of vehicles in a crash, wet road surface, vehicle type, driver age, driver gender, driver seatbelt use and driver alcohol or drug involvement. Among these variables, road curve, functional and disabled vehicle damage in crash, single-vehicle crashes, female drivers, senior drivers, motorcycles and driver alcohol or drug involvement tend to increase the odds of drivers being incapably injured or killed in rural interstate crashes, while wet road surface, male drivers and driver seatbelt use are more likely to decrease the probability of severe driver injuries. The developed methodology and estimation results provide insightful understanding of the internal mechanism of rural interstate crashes and beneficial references for developing effective countermeasures for rural interstate crash prevention.     

Modeling of relative collision safety including driver characteristics

We propose a new mathematical model for relative collision safety in cars. Our present research is restricted to head-on crashes between two cars and we try to determine how much of the injury risk in a crash depends on the car make. Previous work shows that a person’s age and sex influence the injury risk in accidents that are otherwise similar. To explore the relative risks between different car makes we build a model where we let the car mass, change of speed, design of the car and the driver’s age and sex explain the injury outcome in the crashes. The mathematical model we use is a birth process where the states correspond to the injury classes. A database containing police reported traffic accidents and hospital information is used to explore the relationships in our model. Different models are compared and the “best” model is chosen by a likelihood ratio test. The estimated relative risks compensated for the driver’s age and sex are compared to the relative risks with the driver population included. The uncertainties of the different estimates are studied by a bootstrap analysis.     

Factors influential in making an injury severity difference to older drivers involved in fixed object-passenger car crashes

To identify factors influencing severity of injury to older drivers in fixed object-passenger car crashes, two sets of sequential binary logistic regression models were developed. The dependent variable in one set of models was driver injury severity, whereas for the other it was the crash severity (most severe injury in the crash). For each set of models, crash or injury severity was varied from the least severity level (no injury) to the highest severity level (fatality) and vice versa. The source of data was police crash reports from the state of Florida. The model with the best fitting and highest predictive capability was used to identify the influence of roadway, environmental, vehicle, and driver related factors on severity. Travel speed, restraint device usage, point of impact, use of alcohol and drugs, personal condition, gender, whether the driver is at fault, urban/rural nature and grade/curve existence at the crash location were identified as the important factors for making an injury severity difference to older drivers involved in single vehicle crashes.     

Pooling data from fatality analysis reporting system (FARS) and generalized estimates system (GES) to explore the continuum of injury severity spectrum

Fatality Analysis Reporting System (FARS) and Generalized Estimates System (GES) data are most commonly used datasets to examine motor vehicle occupant injury severity in the United States (US). The FARS dataset focuses exclusively on fatal crashes, but provides detailed information on the continuum of fatality (a spectrum ranging from a death occurring within thirty days of the crash up to instantaneous death). While such data is beneficial for understanding fatal crashes, it inherently excludes crashes without fatalities. Hence, the exogenous factors identified as critical in contributing (or reducing) to fatality in the FARS data might possibly offer different effects on non-fatal crash severity levels when a truly random sample of crashes is considered. The GES data fills this gap by compiling data on a sample of roadway crashes involving all possible severity consequences providing a more representative sample of traffic crashes in the US. FARS data provides a continuous timeline of the fatal occurrences from the time to crash – as opposed to considering all fatalities to be the same. This allows an analysis of the survival time of victims before their death. The GES, on the other hand, does not offer such detailed information except identifying who died in the crash. The challenge in obtaining representative estimates for the crash population is the lack of readily available “appropriate” data that contains information available in both GES and FARS datasets. One way to address this issue is to replace the fatal crashes in the GES data with fatal crashes from FARS data thus augmenting the GES data sample with a very refined categorization of fatal crashes. The sample thus formed, if statistically valid, will provide us with a reasonable representation of the crash population.This paper focuses on developing a framework for pooling of data from FARS and GES data. The validation of the pooled sample against the original GES sample (unpooled sample) is carried out through two methods: (1) univariate sample comparison and (2) econometric model parameter estimate comparison. The validation exercise indicates that parameter estimates obtained using the pooled data model closely resemble the parameter estimates obtained using the unpooled data. After we confirm that the differences in model estimates obtained using the pooled and unpooled data are within an acceptable margin, we also simultaneously examine the whole spectrum of injury severity on an eleven point ordinal severity scale – no injury, minor injury, severe injury, incapacitating injury, and 7 refined categories of fatalities ranging from fatality after 30 days to instant death – using a nationally representative pooled dataset. The model estimates are augmented by conducting elasticity analysis to illustrate the applicability of the proposed framework.     

The role of driver sleepiness in car crashes: a systematic review of epidemiological studies

To assess the available evidence for a causal role of driver sleepiness in car crashes or car crash injury, and to quantify the effect, a systematic review of the international literature was conducted. The review included all studies with a fatigue-related exposure measure, a crash or crash injury outcome measure and a comparison group, regardless of publication status, language or date of the study. Eighteen cross-sectional studies and one case-control study fulfilled the inclusion criteria. The fatigue-related exposures investigated in these studies were sleep disorders (n = 14), shift work (n = 2), sleep deprivation/fragmentation (n = 1), and excessive daytime sleepiness (n = 2). Only one study used an injury outcome measure. Studies were limited in their ability to establish a causal relationship by their design, by biases, and in many cases, by small sample sizes. The better quality cross-sectional studies were suggestive of a positive relationship between fatigue and crash risk, but could not provide reliable estimates of the strength of the association. The case-control study provided moderately strong evidence for an association between sleep apnoea and risk of driver injury, with an adjusted odds ratio of 7.2 (95% confidence interval 2.4-21.8). We conclude that the direct epidemiological evidence for a causal role of fatigue in car crashes is weak, but suggestive of an effect. To estimate the burden of injury due to fatigue-related crashes in the population, information is required from well-designed observational epidemiological studies about the prevalence of fatigue in the car driving population and the size of the risk this confers.     

Evolution of the crashworthiness and aggressivity of the Spanish car fleet

This paper investigates the relationship between a passenger car's year of registration and its crashworthiness and aggressivity in real-world crashes. Crashworthiness is defined as the ability of a car to protect its own occupants, and has been evaluated in single and two-car crashes. Aggressivity is defined as the ability to protect users travelling in other vehicles, and has been evaluated only in two-car crashes. The dependent variable is defined as the proportion of injured drivers who are killed or seriously injured; following previous research, we refer to this magnitude as injury severity. A decrease in the injury severity of a driver is interpreted as an improvement in the crashworthiness of their car; similarly, a decrease in the injury severity of the opponent driver is regarded as an improvement in aggressivity.Data have been extracted from the Spanish Road Accident Database, which contains information on every accident registered by the police in which at least one person was injured. Two types of regression models have been used: logistic regression models in single-car crashes, and generalised estimating equations (GEE) models in two-car crashes. GEE allow to take account of the correlation between the injury severities of drivers involved in the same crash. The independent variables considered have been: year of registration of the subject car (crashworthiness component), year of registration of the opponent car (aggressivity component), and several factors related to road, driver and environment.Our models confirm that crashworthiness has largely improved in two-car crashes: when crashing into the average opponent car, drivers of cars registered before 1985 have a significantly higher probability of being killed or seriously injured than drivers of cars registered in 2000–2005 (odds ratio: 1.80; 95% confidence interval: 1.61; 2.01). In single-car crashes, the improvement in crashworthiness is very slight (odds ratio: 1.04; 95% confidence interval: 0.93; 1.16). On the other hand, we have also found a significant worsening in aggressivity in two-car crashes: the driver of the average car has a significantly lower probability of being killed or seriously injured when crashing into a car registered before 1985, than when crashing into a car registered in 2000–2005 (odds ratio: 0.52; 95% confidence interval: 0.45; 0.60).Our results are consistent with a large amount of previous research that has reported significant improvements in the protection of car occupants. They also add to some recent studies that have found a worsening in the aggressivity of modern cars. This trend may be reflecting the impact of differences in masses and travel speeds, as well as the influence of consumer choices. The precise reasons have to be investigated. Also, the causes have to be found for so large a discrepancy between crashworthiness in single and two-car crashes.