Changing patterns in motor vehicle crash mortality: 1940-1980

Overall motor vehicle mortality rates in the United States varied little between 1940 (26 deaths per 100,000 per year) and 1980 (23 deaths per 100,000 per year). However, the distribution of motor vehicle deaths by age, sex, and road user status has changed considerably. Among the elderly there has been a substantial increase in motor vehicle death rates. The result of these changes, and of the altered age structure of the population, is that the number of years of life lost to motor vehicle crash injury before age 70 has increased 84 percent between 1940 and 1980.     

Age and gender patterns in motor vehicle crash injuries: importance of type of crash and occupant role

To evaluate the interaction of gender, age, type of crash, and occupant role in motor vehicle crash injuries leading to hospitalization, we analyzed 1997 Wisconsin hospital discharge data for patients with primary E-code diagnoses of motor vehicle injuries. The overall ratio of males to females (M/F ratio) hospitalized for motor vehicle crash injuries was 1.33 (95% confidence interval (CI): 1.26-1.41). The M/F ratio varied by type of crash and differed for passengers and drivers. For injuries sustained in collisions between vehicles, the M/F ratio was 0.96 (95% CI: 0.87-1.05); in loss of control accidents the M/F ratio was 1.95 (95% CI: 1.76-2.17). Within each type of crash, the M/F ratio for drivers was similar to that for the entire type; the M/F ratio for passengers was about half of the type total. Expressed as rates of hospitalization per 100,000 people in the general population, hospitalizations of drivers in collisions with another motor vehicle increased steeply in males, but not in females, beginning at about age 70. For drivers in loss of control crashes, male rates exceeded female rates in all age groups, with peaks in the groups 15-24 and 85-89. For passengers, injury rates from collisions with other motor vehicles were greater for females, especially in the elderly, and injury rates from loss of control crashes were similar for both genders, with peaks at 15-24 and 85-94. The higher fatality of men in loss of control motor vehicle crashes, compared to women, suggests an important area for further investigation.     

Comparison of injury mortality risk in motor vehicle crash versus other etiologies

The mortality risk ratio (MRR), a measure of the proportion of people who died that sustained a given injury, is reported to be among the most powerful discriminators of mortality following trauma. The primary aim was to determine whether mechanistic differences exist and are quantifiable when comparing MRR-based injury severity across two broadly defined etiologies (motor vehicle crash (MVC) versus non-MVC) for the clarification of important injury types that have some room for improvement by emergency treatment and vehicle design. All International Classification of Diseases, 9th revision (ICD-9) coded injuries in the National Trauma Data Bank (NTDB) database were stratified into MVC and non-MVC groups and the MRR for each injury was computed within each group. Injuries were classified as 11 different types for MRR comparison between etiologies. Overall, MRRs for specific injuries were 10–18% lower for MVC compared to non-MVC etiologies. MVCs however produced much higher mean MRRs for crushing injuries (0.184 versus 0.072) and internal injuries to the thorax, abdomen, and pelvis (0.200 versus 0.169). Non-MVCs produced much higher MRRs for intracranial injuries (0.199 versus 0.250). Analysis of the top 95% most frequent MVC injuries revealed higher MVC MRR values for 78% of the injuries with MRR ratios indicating an average 50% increase in a given injury's MRR when MVC was the etiology. Addressing the large differences in MRR in between etiologies for identical injuries could provide a reduction in fatalities and may be important to patient triage and vehicle safety design.     

The association between sudden illness and motor vehicle crash mortality and injury among older drivers in NSW, Australia

This exploratory study aims to investigate the associations between sudden illness and the risk of motor vehicle crash-related mortalities and injuries among older drivers aged 60 or above. Information utilised in the study was obtained from police reports of all road traffic accidents that occurred on the roads between 1996 and 2000 in New South Wales, Australia. There were 409 older drivers involved in crashes after becoming ill suddenly. Among these the majority (62.1%) of crashes led to at least one occupant in the vehicle being killed or injured. There is a significant association between sudden illness and crash-related mortality and injury after adjusting for other risk factors. The risk of injury and death is increased by nearly six times (OR = 5.58, 95% CI = 4.54-6.85) for those who suffered a sudden illness while driving when compared to those non-sufferers. These results are discussed in the light of possible preventive strategies and the provision of risk assessment and safety counselling for older people.     

The direct and indirect effects of corruption on motor vehicle crash deaths

Recent empirical research has found that there is an inverted U-shaped or Kuznets relationship between income and motor vehicle crash (MVC) deaths, such that MVC deaths increase as national income increases and decrease after reaching a critical level. Corruption has been identified as one of the underlying factors that could affect this relationship, primarily by undermining institutional development and effective enforcement schemes. The total effect of corruption can be decomposed into two components, a direct and an indirect effect. The direct effect measures the immediate impact of corruption on MVC deaths by undermining effective enforcement and regulations, while the indirect effect captures the impact of corruption on hindering increases in per capita income and the consequent impact of reduced income on MVC deaths. By influencing economic growth, corruption can lead to an increase or decrease in MVC deaths depending on the income level. Using data from 60 countries between 1982 and 2003, these effects are estimated using linear panel and fixed effects negative binomial models. The estimation results suggest that corruption has different direct effects for less developed and highly developed countries. It has a negative (decreasing) effect on MVC deaths for less developed countries and a positive (increasing) effect on MVC deaths for highly developed countries. For highly developed countries, the total effect is positive at lower per capita income levels, but decreases with per capita income and becomes negative at per capita income levels of about US$ 38,248. For less developed countries, the total effect is negative within the sample range and decreases with increased per capita income. In summary, the results of this study suggest that reduction of corruption is likely a necessary condition to effectively tackle road safety problems.     

Crossing county lines: the impact of crash location and driver's residence on motor vehicle crash fatality

INTRODUCTION: Studies have demonstrated that the fatality risk for motor vehicle crashes (MVCs) is higher in rural than urban areas. The purpose of this study was to quantify the risk of a fatal outcome associated with a crash by the urban/rural classification of the driver's county of residence and the county of crash before and after adjusting for potentially confounding factors. METHODS: County of crash and driver's county of residence were classified as urban or rural for 514,648 Utah crash participants. Multivariate regression analysis was used to assess the impact of rural versus urban crash location on fatality outcomes for both urban and rural drivers. RESULTS: Before adjusting for confounding factors the relative risk of fatality in a rural versus urban crash was 9.7 (95% CI: 8.0-11.7) for urban drivers and their passengers compared to 1.8 (95% CI: 1.3-2.6) for rural residents. Adjustment for behavioral, road, and crash characteristics reduced risk estimates to 2.8 (95% CI: 2.2-3.5) and 1.2 (95% CI: 0.8-1.7), respectively. CONCLUSION: Urban and rural drivers may have distinct risk factors for MVC fatality in rural areas. Interventions to reduce the risk of fatality in rural areas should evaluate the needs of both urban and rural drivers.     

Under-reporting of motor vehicle traffic crash victims in New Zealand

Our aim was to ascertain the extent of under-reporting of seriously injured motor vehicle traffic crash victims, as recorded by police in New Zealand, and to what extent this coverage was biased by crash, injury, demographic, and geographic factors. Hospital data and police records were linked using probabilistic methods. During 1995, less than two-thirds of all hospitalised vehicle occupant traffic crash victims were recorded by the police. Reporting rates varied significantly by age, injury severity, length of stay in hospital, month of crash, number of vehicles involved, whether or not a collision occurred, and geographic region, but not by gender, ethnicity or day of the week of the crash. Those using these police files for prioritization, resource allocation and evaluation purposes need to be aware of the extent and nature of these biases contained within these databases.     

Spatial analysis of alcohol-related motor vehicle crash injuries in southeastern Michigan

Temporal, behavioral and social risk factors that affect injuries resulting from alcohol-related motor vehicle crashes have been characterized in previous research. Much less is known about spatial patterns and environmental associations of alcohol-related motor vehicle crashes. The aim of this study was to evaluate geographic patterns of alcohol-related motor vehicle crashes and to determine if locations of alcohol outlets are associated with those crashes. In addition, we sought to demonstrate the value of integrating spatial and traditional statistical techniques in the analysis of this preventable public health risk. The study design was a cross-sectional analysis of individual-level blood alcohol content, traffic report information, census block group data, and alcohol distribution outlets. Besag and Newell’s spatial analysis and traditional logistic regression both indicated that areas of low population density had more alcohol-related motor vehicle crashes than expected (P < 0.05). There was no significant association between alcohol outlets and alcohol-related motor vehicle crashes using distance analyses, logistic regression, and Chi-square. Differences in environmental or behavioral factors characteristic of areas of low population density may be responsible for the higher proportion of alcohol-related crashes occurring in these areas.     

The use of generalized estimating equations in the analysis of motor vehicle crash data

The purpose of this study was to determine if it is necessary to use generalized estimating equations (GEEs) in the analysis of seat belt effectiveness in preventing injuries in motor vehicle crashes. The 1992 Utah crash dataset was used, excluding crash participants where seat belt use was not appropriate (n=93,633). The model used in the 1996 Report to Congress [Report to congress on benefits of safety belts and motorcycle helmets, based on data from the Crash Outcome Data Evaluation System (CODES). National Center for Statistics and Analysis, NHTSA, Washington, DC, February 1996] was analyzed for all occupants with logistic regression, one level of nesting (occupants within crashes), and two levels of nesting (occupants within vehicles within crashes) to compare the use of GEEs with logistic regression. When using one level of nesting compared to logistic regression, 13 of 16 variance estimates changed more than 10%, and eight of 16 parameter estimates changed more than 10%. In addition, three of the independent variables changed from significant to insignificant (alpha=0.05). With the use of two levels of nesting, two of 16 variance estimates and three of 16 parameter estimates changed more than 10% from the variance and parameter estimates in one level of nesting. One of the independent variables changed from insignificant to significant (alpha=0.05) in the two levels of nesting model; therefore, only two of the independent variables changed from significant to insignificant when the logistic regression model was compared to the two levels of nesting model. The odds ratio of seat belt effectiveness in preventing injuries was 12% lower when a one-level nested model was used. Based on these results, we stress the need to use a nested model and GEEs when analyzing motor vehicle crash data.     

On the nature of over-dispersion in motor vehicle crash prediction models

Statistical modeling of traffic crashes has been of interest to researchers for decades. Over the most recent decade many crash models have accounted for extra-variation in crash counts--variation over and above that accounted for by the Poisson density. The extra--variation--or dispersion--is theorized to capture unaccounted for variation in crashes across sites. The majority of studies have assumed fixed dispersion parameters in over-dispersed crash models--tantamount to assuming that unaccounted for variation is proportional to the expected crash count. Miaou and Lord [Miaou, S.P., Lord, D., 2003. Modeling traffic crash-flow relationships for intersections: dispersion parameter, functional form, and Bayes versus empirical Bayes methods. Transport. Res. Rec. 1840, 31-40] challenged the fixed dispersion parameter assumption, and examined various dispersion parameter relationships when modeling urban signalized intersection accidents in Toronto. They suggested that further work is needed to determine the appropriateness of the findings for rural as well as other intersection types, to corroborate their findings, and to explore alternative dispersion functions. This study builds upon the work of Miaou and Lord, with exploration of additional dispersion functions, the use of an independent data set, and presents an opportunity to corroborate their findings. Data from Georgia are used in this study. A Bayesian modeling approach with non-informative priors is adopted, using sampling-based estimation via Markov Chain Monte Carlo (MCMC) and the Gibbs sampler. A total of eight model specifications were developed; four of them employed traffic flows as explanatory factors in mean structure while the remainder of them included geometric factors in addition to major and minor road traffic flows. The models were compared and contrasted using the significance of coefficients, standard deviance, chi-square goodness-of-fit, and deviance information criteria (DIC) statistics. The findings indicate that the modeling of the dispersion parameter, which essentially explains the extra-variance structure, depends greatly on how the mean structure is modeled. In the presence of a well-defined mean function, the extra-variance structure generally becomes insignificant, i.e. the variance structure is a simple function of the mean. It appears that extra-variation is a function of covariates when the mean structure (expected crash count) is poorly specified and suffers from omitted variables. In contrast, when sufficient explanatory variables are used to model the mean (expected crash count), extra-Poisson variation is not significantly related to these variables. If these results are generalizable, they suggest that model specification may be improved by testing extra-variation functions for significance. They also suggest that known influences of expected crash counts are likely to be different than factors that might help to explain unaccounted for variation in crashes across sites.     

A laboratory evaluation of a low cost motor vehicle crash recorder

The need for a low cost vehicle crash recorder that would provide a measure of crash severity for a large number of vehicle crashes has long been recognized. A simple viscous damped mechanical crash recorder has been developed which measures vehicle velocity change, ΔV, as the recorded parameter. Prototypes have been evaluated in a laboratory test program which subjected the recorders to a series of tests under various crash speeds and durations that might be expected in the real world. The results of this evaluation indicate that the prototype, low cost crash recorders have great potential for large scale application in motor vehicles.     

The influence of vehicle damage on injury severity of drivers in head-on motor vehicle crashes

Data from crashes investigated through the Crash Injury Research and Engineering Network (CIREN) Program were used to assess differences in injury patterns, severity, and sources for drivers, protected by safety belts and deploying steering wheel air bags, in head-on frontal impacts. We studied whether exterior vehicle damage with a different distribution (wide vs. narrow) across the front vehicle plane influenced injury characteristics. Drivers from both impact types were similar on the basis of demographic characteristics (except age), restraint use, and vehicle characteristics. There were significant differences in the type of object contacted and intrusion into the passenger compartment at the driver's seat location. The mean delta V (based on the kilometers per hour change in velocity during the impact) was similar for drivers in both (wide vs. narrow) impact types. There were no significant differences in injury patterns and sources except that drivers in wide impacts were almost 4 times more likely (odds ratio (OR)=3.81, 95% confidence limits (CL) 1.26, 11.5) to have an abbreviated injury scale (AIS) 3 serious or greater severity head injury. Adjusted odds ratios showed that drivers in wide impacts were less likely (OR=0.54, 95% CI 0.37, 0.79) to have severe injury (based on injury severity score (ISS)>25) when controlling for intrusion, vehicle body type, vehicle curb weight, age, proper safety belt use, and delta V. Drivers with intrusion into their position or who were driving a passenger vehicle were almost twice more likely to have severe injury, regardless of whether the frontal plane damage distribution was wide or narrow. Our study supports that the type of damage distribution across the frontal plane may be an important crash characteristic to consider when studying drivers injured in head-on motor vehicle crashes.     

Multivariate crash modeling for motor vehicle and non-motorized modes at the macroscopic level

Macroscopic traffic crash analyses have been conducted to incorporate traffic safety into long-term transportation planning. This study aims at developing a multivariate Poisson lognormal conditional autoregressive model at the macroscopic level for crashes by different transportation modes such as motor vehicle, bicycle, and pedestrian crashes. Many previous studies have shown the presence of common unobserved factors across different crash types. Thus, it was expected that adopting multivariate model structure would show a better modeling performance since it can capture shared unobserved features across various types. The multivariate model and univariate model were estimated based on traffic analysis zones (TAZs) and compared. It was found that the multivariate model significantly outperforms the univariate model. It is expected that the findings from this study can contribute to more reliable traffic crash modeling, especially when focusing on different modes. Also, variables that are found significant for each mode can be used to guide traffic safety policy decision makers to allocate resources more efficiently for the zones with higher risk of a particular transportation mode.     

Indexing crash worthiness and crash aggressivity by vehicle type

Crash aggressivity (CA), along with conventional crash worthiness (CW), has been recently studied to deal with the crash incompatibility between vehicles on roads. Clearly, injury severity depends on the attacking ability of striking vehicle as well as the protective ability of struck vehicle. This study proposes a systematic crash-based approach to index CA and CW of various vehicles. The approach deviates from existing methods in three aspects: (a) an explicit definition and specification in the model for CW and CA; (b) Bayesian hierarchical analysis to account for the crash-vehicle two-level data structure; (c) a five-level ordinal model to explicitly consider all levels of crash severity. The case study on major vehicle types illustrated the method and confirmed the consistency of results with previous studies. Both crash worthiness and crash aggressivity significantly vary by vehicle types, in which we identified the dominating effect of vehicle mass, and also highlighted the extraordinary aggressivity of Light Trucks and Vans (LTVs). While it was not surprising to identify least CA and CW of motorcycles, buses were unconventionally found to be less aggressive than other motor vehicles. The method proposed in this research is applicable to detailed crash-based vehicle inspection and evaluation.     

Urban-rural differences in motor vehicle crash fatality and hospitalization rates among children and youth

This population-based study examined motor vehicle crash hospitalization rates and death rates among children and youth in rural and urban areas of the province of Alberta, Canada. Using police report data (1997-2002, inclusive), average annual motor vehicle crash hospitalization and death rates among those 0-19 years of age were calculated for rural and urban regions. Across all age and sex strata examined, both the hospitalization and the fatality rates were significantly higher in rural compared with urban regions. After adjusting for age, sex and calendar year, the relative risk of a motor vehicle crash hospitalization (rural versus urban) was 3.0 (95% CI: 2.8, 3.2). After adjusting for age, sex and calendar year, the relative risk of a motor vehicle crash fatality was 5.4 (95% CI: 4.2, 6.9). Motor vehicle crash injury hospitalization and fatality rates among children and youth in the province of Alberta are considerably higher in rural areas compared with urban areas. There is a need to identify social, demographic and environmental driving hazards associated with the rural environment.     

The epidemiology of traumatic injury-related fetal mortality in Pennsylvania, 1995-1997: the role of motor vehicle crashes.

METHODS: Rates and causes of traumatic injury-related fetal deaths in Pennsylvania were determined from a manual review of all fetal death certificates filed from 1995 to 1997 (7,131 cases). RESULTS: Thirty one traumatic injury cases were identified (6.5/100,000 live births). Most cases (94%) could be identified from the diagnosis code of 760.5 (maternal injury) and 87% contained narratives indicating specific injury mechanisms. Motor vehicles were the leading cause of injury (81%). Placental separation was the leading diagnosis (42%). CONCLUSIONS: The ICD-9-CM code 760.5 appears to be a specific indicator of traumatic fetal death, though it is not known how sensitive an indicator it is. Though not usually E-coded, the death certificates contained enough information to allow ascertainment of injury mechanism. These are very conservative estimates of the burden of the problem. The major role that motor vehicle injuries have on reported traumatic fetal injury deaths was shown and a significant new challenge for child passenger safety advocates is indicated.     

A small area study of motor vehicle crash fatalities in Alberta,Canada

This study examined the small area variation in motor vehicle crash fatality rates in the province of Alberta, Canada. Motor vehicle crash fatality rates per 100000 population (1995-1997, inclusive) were determined for five geographic areas in the province. The rates showed substantial, statistically significant variation across areas, with fatality rates lowest in the urban areas of Calgary and Edmonton, and highest in the rural areas (south, central, and northern Alberta). Examination of area-level predictors-population density, impaired driving citation rates, education level, unemployment levels, and ethnicity-showed that population density and impaired driving rates were associated with motor vehicle crash fatality rates. There was a five-fold difference in annual motor vehicle crash fatality rates between rural (22.9/100000) and urban areas (4.4/100000), whereas annual impaired driving rates were around 1.8% in rural areas, compared with 0.6% in urban areas. Because of multicollinearity problems, it was not possible to estimate a multivariable Poisson regression model. In conclusion, rural areas in the province of Alberta demonstrate a significantly higher motor vehicle crash fatality rate, compared with urban areas.     

A comparison of diagnostic related group length of stay outliers: motor vehicle crash versus penetrating injuries

The incidence and degree to which patients injured by motor vehicle crashes (MVCs) and penetrating wounds remain in the hospital beyond the diagnostic related group (DRG) mean length of stay (LOS) are compared. During a 12-month period, records for consecutive patients admitted to eight hospitals (including three trauma centers) were studied. Patients aged 13 or younger, staying less than 48 hours, or with major burns or only distal fractures were excluded. In that time, 2,914 patients were eligible. Twenty percent of injuries were penetrating; 27% were MVC injuries; and, 53% were blunt injuries from other causes. Patients injured in MVCs and with penetrating injuries were compared with respect to mean LOS, incidence of DRG outliers, number of hospital days beyond the DRG mean LOS, and demographic variables. Patients injured in MVCs had a greater proportion of DRG LOS outliers and higher mean numbers of DRG excess days than did patients with penetrating wounds (p less than 0.01, for both). Injuries were distributed among relatively more DRGs for MVC patients. The DRG scheme may lack sufficient attention to factors more likely to affect MVC patients, such as multiplicity of injuries, incidence of CNS injuries, ICU requirements, and older age. In structuring more appropriate reimbursement for trauma care, special attention must be paid to patients injured in MVCs.     

Predicting severe head injury after light motor vehicle crashes: implications for automatic crash notification systems

Motor vehicle crashes (MVC) are a leading public health problem. Improving notification times and the ability to predict which crashes will involve severe injuries may improve trauma system utilization. This study was undertaken to develop and validate a model to predict severe head injury following MVC using information readily incorporated into an automatic crash notification system. A cross-sectional study with derivation and validation sets was performed. The cohort was drawn from drivers of vehicles involved in MVC obtained from the National Automotive Sampling System (NASS). Independent multivariable predictors of severe head injury were identified. The model was able to stratify drivers according to their risk of severe head injury indicating its validity. The areas under the receiver-operating characteristic (ROC) curves were 0.7928 in the derivation set and 0.7940 in the validation set. We have developed a prediction model for head injury in MVC. As the development of automatic crash notification systems improves, models such as this one will be necessary to permit triage of what would be an overwhelming increase in crash notifications to pre-hospital responders.