Accident Analysis and Prevention 117 (2018) 98–105
Contents lists available at ScienceDirect
Accident Analysis and Prevention
journal homepage: www.elsevier.com/locate/aap
Accident analysis to support the development of strategies for the
prevention of brain injuries in car crashes
T
⁎
Jacobo Antona-Makoshia, , Koji Mikamia, Mats Lindkvistb, Johan Davidssonc, Sylvia Schickd
a
Japan Automobile Research Institute, 2530 Karima, Tsukuba, Ibaraki, 305-0822, Japan
Umeå University, SE-901 87, Umeå, Sweden
c
Chalmers University, SE-412 96, Gothenburg, Sweden
d
Ludwig-Maximilians-Universität LMU, Munich, Germany
b
A R T I C LE I N FO
A B S T R A C T
Keywords:
NASS-CDS
Traumatic brain injuries
Concussion
Subdural haemorrhage
Sex
Age
This study estimated the frequency and risk of Moderate-to-Maximal traumatic brain injuries sustained by occupants in motor vehicle crashes in the US. National Automotive Sampling System - Crashworthiness Data
System crashes that occurred in years 2001–2015 with light vehicles produced 2001 or later were incorporated
in the study. Crash type, crash severity, car model year, belt usage and occupant age and sex were controlled for
in the analysis. The results showed that Moderate concussions account for 79% of all MAISbrain2+ injuries. Belted
occupants were at lower risks than unbelted occupants for most brain injury categories, including concussions.
After controlling for the effects of age and crash severity, belted female occupants involved in frontal crashes
were estimated to be 1.5 times more likely to sustain a concussion than male occupants in similar conditions.
Belted elderly occupants were found to be at 10.5 and 8 times higher risks for sub-dural haemorrhages than nonelderly belted occupants in frontal and side crashes, respectively. Adopted occupant protection strategies appear
to be insufficient to achieve significant decreases in risk of both life-threatening brain injuries and concussions
for all car occupants. Further effort to develop occupant and injury specific strategies for the prevention of brain
injuries are needed. This study suggests that these strategies may consider prioritization of life-threatening brain
vasculature injuries, particularly in elderly occupants, and concussion injuries, particularly in female occupants.
1. Introduction
Traumatic Brain Injuries (TBIs) account for about half of the 1.3
million annual traffic related deaths and the 50 million traffic related
injuries worldwide (World Health Organization, 2013). Vehicle occupants comprise the largest group of road traffic deaths by road user type
in most high-income countries (World Health Organization, 2013). TBIs
are the main cause of death and severe injuries amongst most vehicle
crash types and population groups; young adults and the elderly are at
higher risks than mature adults (Bruns and Hauser, 2003; Bener et al.,
2010) and females are at higher risks than males (Bose et al., 2011).
Comprehensive estimations of costs per body region injured in vehicle
crashes in the US, including medical costs, emergency services, lost
work wages and loss of quality of life, among others, point at TBIs as the
second most costly after spinal cord injuries (Zaloshnja et al., 2004).
Fatal TBIs in Motor Vehicle Crashes (MVCs) produce a total estimated
annual comprehensive societal cost in the US of US$ 62.8 billion (Eigen
and Martin, 2005). Despite the ever-improving vehicle occupant
⁎
protection, TBI prevention strategies must still be assigned top priority.
Occupant injury prevention strategies target at the reduction of the
risk of fatal and severe head injuries in high severity crash tests.
Unfortunately, these strategies have proven insufficient to achieve
significant reductions of these TBIs risks (Eigen and Martin, 2005;
Takhounts et al., 2013). Non-life threatening TBIs, which are mainly
concussions, have gained less attention in the development of the
prevention strategies. Still, recent traffic data analysis has shown that
Moderate-to-Serious concussions account for 60% of all head injuries in
MVCs (Viano and Parenteau, 2015) and several studies have shown that
the consequences of a concussion often cause fatigue, sleep disorders,
sequelae, long term pain, vertigo, mood disorders and depression
(Carroll et al., 2014). As many as 30% of concussion victims may suffer
persistent cognitive, physical or psychosocial impairment (Carroll et al.,
2014). Hence recent studies call for updated head injury prevention
strategies that reduce concussion injuries, in addition to fatal and severe head injuries. For the development of these strategies, understanding the real-world crash scenarios and the occupant factors of
Corresponding author.
E-mail addresses: ajacobo@jari.or.jp (J. Antona-Makoshi), kmikami@jari.or.jp (K. Mikami), mats.lindkvist@umu.se (M. Lindkvist), johan.davidsson@chalmers.se (J. Davidsson),
sylvia.schick@med.uni-muenchen.de (S. Schick).
https://doi.org/10.1016/j.aap.2018.04.009
Received 30 June 2017; Received in revised form 7 March 2018; Accepted 7 April 2018
Available online 19 April 2018
0001-4575/ © 2018 Elsevier Ltd. All rights reserved.
Accident Analysis and Prevention 117 (2018) 98–105
J. Antona-Makoshi et al.
database, SAS Enterprise Guide version 7.1 (SAS Institute Inc., North
Carolina, USA) software was utilized.
Detailed descriptions of the inclusion criteria, the injury classification and the statistical methods applied follow.
relevance for specific TBIs is required.
The National Automotive Sampling System - Crashworthiness Data
System (NASS-CDS) is a publicly available US nationwide MVC data
collection program developed and maintained by the National Highway
Traffic Safety Administration (NHTSA) (Zhang and Chen, 2013). NASSCDS is a probability sample survey of towed-away and police-reported
MVCs with a complex sample design that allows for national estimates
on crash, vehicle, and occupant characteristics (Pintar et al., 2000;
Zhang and Chen, 2013). The main purpose of the NASS-CDS is to
support the development of traffic safety countermeasures. A number of
studies have used the NASS-CDS to analyze head and brain injuries in
MVCs. The risk of Severe head injuries has been shown to increase with
crash severity (Talmor et al., 2006) and with occupant’s age (Ridella
et al., 2012). The rate of extra-axial bleeding injuries has been shown to
increase with age in 1993–2007 crashes (Mallory, 2010). A former
study analyzed 1991 to 1998 crashes to conclude that the use of seatbelts and airbags was associated with reduced frequencies of skull and
brain injuries of moderate or worse severity in front row occupants
(Pintar et al., 2000). More recently, an analysis of crashes from years
1994 to 2011 has estimated the frequency and risks of concussion injuries (Viano and Parenteau, 2015). The study showed that the risk of
concussions is reduced with belt use and that the majority of these
injuries occur at crash severity speed changes of less than 40 km/h
(Viano and Parenteau, 2015). None of the previous studies that incorporated Moderate brain injuries (Pintar et al., 2000; Viano and
Parenteau, 2015) provided an analysis of the demographics of concussion victims or a statistical analysis of how different crash, vehicle
and occupant factors affect risks for concussion injuries.
The aim of this study is to estimate the frequency and risk of all
Moderate-to-Maximal (AIS2+) brain injuries sustained by occupants in
reported MVCs, and to analyze the data considering crash year, crash
type, crash severity, car model year, belt usage and the victim’s age and
sex.
2.1. Inclusion criteria
The following inclusion criteria were applied:
• Crash year 2001–2015
• Vehicle model year 2001–2015
• Light vehicles (passenger cars, pick-ups and mini-vans)
• Non-ejected occupants
• Occupant age 15 or higher
• Occupants with known injury status or fatality
Light vehicles was defined with the NASS variable VEHICLE MODEL
(Light vehicles when variable was between 0–490). Occupant ejection
was defined with the NASS variable EJECTION (Non-Ejected when
variable was 0). Occupants with reported injury status are those assigned a MAIS between 0 and 6 or Fatality. Crash types were defined in
five categories with the NASS variables GAD1 (highest area of vehicle
damage) and ROLLOVER (Frontal if GAD1 = F and ROLLOVER < = 0;
Side if GAD1 = L or R and ROLLOVER < = 0; Rear if GAD1 = B and
ROLLOVER < = 0; Rollover if ROLLOVER > 0; and Other for occupants in crash types not included in any of the previous categories).
Belted occupants were defined as those in which the investigator
found evidence of belt use and that the used belt was of 3-point belt
type (shoulder belt and a lap belt in combination) (Belted when variable MANUSE = 4). All other occupants, including those not using any
belt, those with an inoperative belt, those only using a shoulder belt,
those only using a lap belt, or those using a belt which type could not be
confirmed, were grouped in a category and denoted as Unbelted.
Occupant's sex was defined using the variable SEX (Male if SEX = 1 and
Female if SEX = 2–6). Change in velocity (DVTOTAL) of the occupant's
vehicle was used to estimate the crash severity.
2. Methods
NASS-CDS data from years 2001–2015 were analyzed to estimate
the frequency and risk of TBIs sustained by occupants in MVCs. First, a
descriptive analysis was conducted to estimate the frequency and risks
for different TBI categories by crash type, crash year and belt use.
Second, logistic regression was applied to model the effects of car model
year, crash severity, occupant's sex and age on the odds of sustaining
the most relevant injuries identified.
Injuries in NASS-CDS are documented by NASS investigators according to the Abbreviated Injury Scale (AIS)© for which they are
previously trained and certified. The AIS is an anatomically based and
consensus derived scoring system. Every seven digit AIS code consists of
a six digit injury descriptor and an AIS severity code that classifies an
individual injury by body region according to its relative severity on a
six point ordinal scale (Gennarelli and Wodzin, 2006). In this scale,
AIS1 stands for Minor injuries, AIS2 for Moderate, AIS3 for Serious, AIS4
for Severe, AIS5 for Critical, and AIS6 for Maximal (currently untreatable) injuries. The NASS investigators follow up their on-site investigations by interviewing crash victims to preliminary determine the
nature and severity of injuries. Thereafter, the investigators get access
to hospital records that are provided by the medical community. These
records are contrasted with the preliminary investigator observations
and become the primary source of data to judge the nature and severity
of injuries.
NASS-CDS utilizes several parallel versions of AIS. In this study,
NASS-CDS cases from years 2001 to 2015 with injury codes according
to AIS-90/98 were downloaded from the NHTSA resources and merged
into a single database of occupants. Crash, vehicle and injury data relevant for this study were assigned to their corresponding occupants. To
build the occupants database, SPSS Statistics version 24 (IBM
Corporation, New York, USA) software was utilized. To analyze the
2.2. Injury categories
All AIS2+ brain injuries were categorized based on seven-digit AIS
codes. A broad description of the brain injury categories that comprised
the main focus of this study is provided below. A table with the injury
category assigned to each seven-digit AIS code is provided in the appendix (Table A8)
• Concussions: AIS2-3 injuries documented as Cerebral Concussion, as
•
•
•
•
•
•
•
99
Length of Unconsciousness, as Level of Consciousness, or as
Lethargic, Stuporous, or Obtunded post resuscitation.
Contusions: AIS3-5 injuries documented as contusions in the cerebrum or the cerebellum.
Diffuse Axonal Injuries: AIS4-5 injuries documented as white matter
shearing in the cerebrum or the cerebellum (but not in the brainstem), as Length of Unconsciousness, or as Level of Consciousness.
Epidural Haemorrhages: AIS4-5 injuries reported as epidural or
extradural haematoma/haemorrhage in the cerebrum or the cerebellum.
Subdural Haemorrhages: AIS4-5 injuries reported as subdural haematoma/haemorrhage in the cerebrum or the cerebellum.
Sub-Arachnoid Haemorrhages: AIS3 injuries reported as subarachnoid or subpial haemorrhage in the cerebrum or the cerebellum.
Intracranial Haemorrhages: AIS4-5 injuries reported as intracerebral, intracerebellar or intraventricular haematoma/haemorrhage.
Brainstem Injuries: AIS5-6 injuries reported as injuries to the
brainstem, including compression, contusion, DAI, infarction,
Accident Analysis and Prevention 117 (2018) 98–105
J. Antona-Makoshi et al.
•
the sample period, their percentage distribution, the mean age, and the
belt use rates by occupant's sex and three age groups. Nearly half of the
occupants were females. For both sexes, the Young occupant group (age
15–34) dominated the sample, followed by the Adult (age 35 to 64) and
the Elderly (age 65+) occupant groups. Overall belt use rate was 0.79,
with lower rates in males (0.76) than in females (0.82). The Young
occupant groups accounted for the lowest belt rates (0.75 in males, 0.80
in females). These rates increased for Adults (0.76 in males, 0.83 in
females) and became the highest and equal in Elderlies (0.87 for both
sexes).
laceration, massive destruction, penetration and no further specified
(NFS).
BrainOthers: all AIS3-5 injuries to the brain not included in any of
the above categories. Examples of injuries included in this group are
pituitary injuries, swelling, infarction, penetration, and injuries in
the cerebrum or the cerebellum NFS.
2.3. Descriptive analysis
Frequency of a particular brain injury was defined as the number of
occupants who sustained that particular injury as their highest severity
brain injury (MAISbrain). Whenever an occupant had two or more injuries of the same and highest severity and from distinct categories, that
occupant was counted once for each of the distinct injury categories.
The risk that an occupant experienced a particular brain injury was
calculated as the number of occupants with that particular MAISbrain
injury divided by the number of occupants with known injury status.
National estimates based on weighted data and the standard errors of
these estimates were calculated by using the SAS SURVEYFREQ procedure that accounts for stratification (PSUSTRAT), clustering (PSU),
and sampling weight (RATWGT) variables. 95 % confidence limits of
the estimated frequencies and risks were calculated by multiplying the
standard errors by a factor of 1.96 and are indicated throughout the
manuscript following the ‘ ± ’ sign. In addition to MAISbrain injuries,
since concussions may accompany other injuries of higher severity and
thus these would be classified as the MAISbrain, the number of occupants
who sustained a particular injury, regardless if the injury was of highest
severity, was also estimated (AISbrain).
3.1. Brain injury category by injury severity
Table 2 shows the frequency ± 95% confidence limits for
MAISbrain2+ injuries by injury category and severity. Out of the
20,395,953 occupants contained in the sample, a total of
367,203 ± 148,527 occupants sustained a AIS2-3 concussion as their
highest severity brain injury, which accounts for 79.3% of all occupants
with MAISbrain2+ injuries. The large majority of these concussions
(99.4%) were of AIS2 severity. A total of 10,039 ± 7922 occupants
sustained a AIS3-5 contusion. The majority of these contusions (93.4%)
were of AIS3 severity. Also at the AIS3 level, 17,823 ± 8226 occupants
sustained a SAH. For AIS4-5 brain injuries, the most frequent injury
category was SDH (22,308 ± 7283 occupants), followed by ICH
(16,797 ± 11,539 occupants), DAI (6182 ± 4787 occupants), and
EDH (3055 ± 962 occupants). AIS5-6 brainstem injuries accounted for
a total of 8665 ± 4281 occupants, of which 3644 ± 1421 were of
AIS6 severity.
2.4. Odds ratio analysis
3.2. Brain injury category by crash type
This part of the study was limited to belted occupants in frontal or
side crashes from which crash severity (delta-V) was available. The
limitation to belted occupants was a deliberate attempt to focus on the
relevant scenarios in which occupants get injured despite using a belt
and that represent the large majority of cases in current and future
expected scenarios (Pickrell, 2017). The SAS SURVEYLOGISTIC procedure was applied to model the effects of car model year, crash severity,
occupant's sex and age on the odds to sustain the most relevant brain
injury categories identified. Each of these injury categories was considered as a dichotomous categorical variable equal to one for occupants with the injury and zero otherwise.
Delta-V was modeled as a continuous variable. Sex and Age were
considered as dichotomous variables (Female vs Male; Elderly 65 + vs
Non-elderly 15–64). Three car model year groups (2001-2004,
2005–2010, 2011-2015) were incorporated according to years in which
changes of safety regulations or assessment programs may have influenced the safety performance of vehicles. The 2001-2004 group represented the first generation of vehicles with depowered airbags. The
2005–2010 group complied with FMVSS301 updates and included vehicle models that were evaluated in the first generations of consumer
ratings such as the IIHS top safety pick awards. The 2011-2015 group
were those subjected to testing according to the enhanced US-NCAP. In
addition to car model year, a variable to control for crash year before or
after 2010 was considered. The incorporation of this variable was an
attempt to improve the understanding of possible effects that the incorporation of the AIS-05/08 version into NASS-CDS may have had on
the likelihood of concussive brain injuries to be documented in the
dataset.
Table 3 shows the frequency and risks ± 95% confidence interval
limits for occupants with MAISbrain2+ injuries by injury category and
crash type. The frequency and risk for concussions were the highest
among all crash types. Rollovers accounted for the highest injury risk
(3.613 ± 1.605%), followed by Side (2.117 ± 1.063%), Rear
(2.098 ± 1.817%) and Frontal crashes (1.965 ± 0.799%). For nonconcussion injuries SDH, SAH and Contusions comprised the three most
common injury categories in frontal and rollover crashes. In side crashes, ICH was the most common category followed by SDH and SAH. In
rear crashes, SDH and SAH were the most common categories. An
analogous table is provided in the appendix (Table A9) with the frequency and risks for occupants with AIS2 + brain injuries.
3.3. Brain injury risk by crash year
Figs. 1 and 2 show raw and weighted risks for MAISbrain3+ injuries
and Concussions, respectively, by crash year. Both figures are presented
as running three year average risk values, as an attempt to mitigate
year-by-year fluctuations inherent to NASS-CDS data (Eigen and
Martin, 2005; Takhounts et al., 2013). For example, risks in year 2003
are calculated as the average of the risks from years 2001 to 2003. For
the period analyzed, weighted MAISbrain3+ injury risk appears to
fluctuate around an approximately constant value. Weighted concussion injury risk appears to initiate a steady increase from year 2006 that
intensifies from year 2010.
3.4. Effect of belt use on brain injury risk
3. Results
Figs. 3 and 4 show the effect of belt use on MAISbrain3+ injury and
on concussion injury risk, respectively. The bars in the figures represent
the 95% confidence intervals. The tables containing the data from
which the figures were constructed are provided in the appendix (Table
A8). Unbelted occupants were at higher risks than belted occupants for
most brain injury categories, including concussions.
Applying the inclusion criteria to the NASS-CDS data resulted in a
dataset consisting of total raw number of 49,855 occupants. Applying
the weighting factors a total of 20,370,955 occupants were obtained.
Table 1 shows the total raw and the weighted numbers of occupants for
100
Accident Analysis and Prevention 117 (2018) 98–105
J. Antona-Makoshi et al.
Table 1
Total sample size, occupant demographics and belt use rate based on 15 years of NASS-CDS data (2001–2015).
Sex
Age Group
Raw No. of Occupants
Weighted No. of Occupants
% of Occupants
Mean Age (Years)
Belt Use Rate
Female
Young 15-34
Adult 35-64
Elderly 65+
Total (F)
12,115
10,053
2,525
24,693
5,239,318
3,960,334
962,131
10,161,783
25.7
19.4
4.7
49.8
23.8
48.1
73.7
38.0
0.80
0.83
0.87
0.82
Male
Young 15-34
Adult 35-64
Elderly 65+
Total (M)
13,214
9,611
2,337
25,162
5,441,293
3,812,942
979,934
10,680,611
26.7
18.7
4.8
50.2
23.9
47.5
73.9
37.5
0.75
0.76
0.87
0.76
49,855
20,395,953
100
37.7
0.79
Total (F + M)
Table 2
Frequency of occupants with MAISbrain2+ injuries by injury category and severity based on 15 years of NASS-CDS weighted data (2001-2015).
MAISbrain 2
MAISbrain 3
MAISbrain 4
MAISbrain 5
MAISbrain 6
TOTAL
Concussion
Contusion
DAI
Brainstem
SDH
SAH
EDH
ICH
BrainOthers
365,183 ± 147,463
–
–
–
–
–
–
–
–
2,020 ± 1064
9381 ± 7,405
–
–
–
17,823 ± 8226
–
–
5054 ± 2,644
–
562 ± 416
2454 ± 3,303
–
16,625 ± 4,953
–
2609 ± 709
10,763 ± 5,766
2381 ± 786
–
96 ± 100
3,728 ± 1485
5021 ± 2,860
5,683 ± 2330
–
446 ± 254
6,034 ± 5772
826 ± 347
–
–
–
3644 ± 1421
–
–
–
–
–
367,203 ± 148,527
10,039 ± 7922
6182 ± 4787
8665 ± 4281
22,308 ± 7283
17,823 ± 8226
3055 ± 962
16,797 ± 11,539
8261 ± 3,777
Total
365,183 ± 147,463
34,278 ± 19,339
35,394 ± 15,933
21,834 ± 13,148
3644 ± 1421
460,333 ± 197,304
Estimate ± 95% confidence limit of the estimate.
indicate that, after adjusting for the effects of age, crash year and deltaV, a belted female occupant involved in a frontal crash with a vehicle
produced 2001 or later is estimated to be 1.5 times more likely to
sustain a concussion than a belted male occupant. In addition, for a
given vehicle model year group and for comparable crashes, concussions were 1.6 times more likely to be reported for crashes occurring
2010 or later as compared to crashes occurring 2009 and earlier.
Analogously, limiting the total sample to belted occupants in side
crashes with available delta-V resulted in a dataset containing
2,389,173 weighted occupants from which 33,144 sustained a concussion. Table 5 presents the results from applying logistic regression to
3.5. Odds ratio analysis for Concussion in frontal and side crashes
Limiting the total weighted sample to belted occupants in frontal
crashes with available delta-V resulted in a dataset containing
4,720,890 occupants from which 55,272 sustained a concussion.
Table 4 summarizes the results from applying logistic regression to the
occurrence of concussion for this subset of frontal crashes. The p-values
in the last column show that delta-V (p = 0.0003), sex (p = 0.0244),
age (p = 0.0037), and crash year 2010+ (p = 0.0134) were significant
contributors, with a 95% confidence level, to predict the probability of
concussion. The odds ratio point estimates in the second column
Table 3
Frequency and risks for occupants with MAISbrain2+ injury by crash type based on 15 years of NASS-CDS weighted data (2001–2015).
Frontal
Side
Rear
Rollover
Others
No. occupants
Known inj. status
No head injury
Concussion
Contusion
DAI
Brainstem
SDH
SAH
EDH
ICH
BrainOthers
9,299,435 ± 2,157,940
8,785,034 ± 2,005,403
182,718 ± 74,337
5,457 ± 7,142
1,297 ± 440
3,291 ± 1,435
8,185 ± 3,256
8,108 ± 4,261
1,068 ± 560
4,939 ± 2,532
2,992 ± 1,189
4,523,767 ± 1,280,147
4,155,377 ± 1,145,167
95,753 ± 48,079
2,450 ± 765
3,145 ± 2783
2,835 ± 1767
7,959 ± 4,157
7,029 ± 5,766
757 ± 378
8,187 ± 6152
3,462 ± 2,413
1,284,847 ± 393,531
1,193,507 ± 364,832
26,954 ± 23,348
(84 ± 140)
(183 ± 321)
(853 ± 1155)
2,733 ± 1,997
1,239 ± 1,056
(47 ± 83)
(166 ± 190)
28 ± 18
898,516 ± 372,355
713,764 ± 275,331
32,467 ± 14,424
1462 ± 1196
1015 ± 926
722 ± 403
1,551 ± 410
1,406 ± 608
963 ± 888
1148 ± 466
780 ± 538
4,389,389 ± 1,276,891
4,258,224 ± 1,242,279
29,311 ± 13,516
(586 ± 633)
543 ± 322
964 ± 520
1,880 ± 941
656 ± 161
(219 ± 265)
(2357 ± 2824)
998 ± 580
Injury Risk (%)
Concussion
Contusion
DAI
Brainstem
SDH
SAH
EDH
ICH
BrainOthers
1.965
0.059
0.014
0.035
0.088
0.087
0.011
0.053
0.032
2.117
0.054
0.070
0.063
0.176
0.155
0.017
0.181
0.077
2.098 ± 1.817
(0.007 ± 0.011)
(0.014 ± 0.025)
(0.066 ± 0.090)
0.213 ± 0.155
0.096 ± 0.082
(0.004 ± 0.006)
(0.013 ± 0.015)
0.002 ± 0.001
3.613
0.163
0.113
0.080
0.173
0.156
0.107
0.128
0.087
0.668 ± 0.308
(0.013 ± 0.014)
0.012 ± 0.007
0.022 ± 0.012
0.043 ± 0.021
0.015 ± 0.004
(0.005 ± 0.006)
(0.054 ± 0.064)
0.023 ± 0.013
±
±
±
±
±
±
±
±
±
0.799
0.077
0.005
0.015
0.035
0.046
0.006
0.027
0.013
±
±
±
±
±
±
±
±
±
1.063
0.017
0.062
0.039
0.092
0.127
0.008
0.136
0.053
±
±
±
±
±
±
±
±
±
1.605
0.133
0.103
0.045
0.046
0.068
0.099
0.052
0.060
Estimate ± 95% confidence limit of the estimates. In parenthesis the estimates for which the 95% confidence limit values were higher than the estimate.
101
Accident Analysis and Prevention 117 (2018) 98–105
J. Antona-Makoshi et al.
Table 4
Odds ratio estimates based on logistic regression analysis of Concussion cases.
Significant (p < 0.05) point estimates marked with an asterisk. (Crash Year
2001–2015, Car Model Year 2001–2015, Non-ejected, Age 15+, Belted, Frontal
crashes).
Fig. 1. MAISbrain3+ injury risk by crash year (Raw risk in red thin continuous
line. Weighted risk in black thick continuous line. 95% confidence corridor
limits in black dotted lines) (For interpretation of the references to colour in this
figure legend, the reader is referred to the web version of this article).
Effect
Estimate
95% Confidence Limits
p-value
delta-V
Sex (Female vs Male)
Elderly (65 + vs 15-64)
Crash Year (2010+ vs 2001/
09)
Model Year (2001/04 vs 2011/
15)
Model Year (2005/10 vs 2011/
15)
1.05*
1.51*
0.49*
1.61*
1.03
1.06
0.32
1.12
1.07
2.14
0.77
2.31
0.0003
0.0244
0.0037
0.0134
1.11
0.43
2.88
0.9059
1.31
0.46
3.71
0.4647
Table 5
Odds ratio estimates based on logistic regression analysis of Concussion cases.
Significant (p < 0.05) point estimates marked with an asterisk. (Crash Year
2001–2015, Car Model Year 2001–2015, Non-ejected, Age 15+, Belted, Side
crashes).
Effect
Estimate
95% Confidence Limits
p-value
delta-V
Sex (Female vs Male)
Elderly (65+ vs 15–64)
Crash Year (2010+ vs 2001/
09)
Model Year (2001/04 vs
2011/15)
Model Year (2005/10 vs
2011/15)
1.08*
1.13
1.96
1.20
1.07
0.49
0.37
0.64
1.09
2.58
10.22
2.25
< 0.0001
0.7599
0.3992
0.5426
3.01*
1.52
5.98
0.0096
2.40
1.18
4.85
0.1885
the occurrence of concussion for this set of side crashes. Only delta-V
(p < 0.0001) and Car Model Year 2011+ (p = 0.009) were significant
at the 95% confidence level.
Fig. 2. Concussion injury risk by crash year (Raw risk in red thin continuous
line. Weighted risk in black thick continuous line. 95% confidence corridor
limits in black dotted lines) (For interpretation of the references to colour in this
figure legend, the reader is referred to the web version of this article).
3.6. Odds ratio analysis for SDH in frontal and side crashes
From the 4,720,890 belted occupants involved in a frontal crash
(and for which delta-V was available), 2583 occupants sustained a SDH.
Table 6 summarizes the results from applying logistic regression to the
occurrence of SDH in this set of frontal crashes. The p-values in the last
column show that, within the variables incorporated, only delta-V and
age group significantly contribute to predict the probability of SDH.
Sex, crash year 2010 or car model year were not significant contributors
to predict the risk of SDH. The odds ratio estimates in the second
column indicate that, after adjusting for the effects of delta-V, a belted
elderly occupant involved in a frontal crash with a vehicle produced
2001 or later is estimated to be 10.5 times more likely to sustain a SDH
than a non-elderly belted adult.
Fig. 3. Effect of belt use on MAISbrain3+ injury risk with 95% confidence intervals.
Table 6
Odds ratio estimates based on logistic regression analysis of SDH cases.
Significant (p < 0.05) point estimates marked with an asterisk. (Crash Year
2001–2015, Car Model Year 2001–2015, Non-ejected, Age 15+, Belted, Frontal
crashes).
Fig. 4. Effect of belt use on Concussion injury risk with 95% confidence intervals.
102
Effect
Estimate
95% Confidence Limits
p-value
delta-V
Sex (Female vs Male)
Elderly (65+ vs 15-64)
Crash Year (2010+ vs 2001/
09)
Model Year (2001/04 vs
2011/15)
Model Year (2005/10 vs
2011/15)
1.08*
1.43
10.56*
0.98
1.054
0.382
3.98
0.342
1.099
5.385
28.023
2.797
< 0.0001
0.5706
0.0001
0.9646
0.76
0.041
14.35
0.7274
1.12
0.101
12.313
0.691
Accident Analysis and Prevention 117 (2018) 98–105
J. Antona-Makoshi et al.
2010). The current study, although less detailed in terms of analysis of
accompanying injuries, is consistent with past findings on age as a main
risk factor of bleeding injuries, particularly SDH. The current study also
suggests that this increased SDH risk in elderly persists in modern vehicles. The overall results of the current and past studies call for the
development of strategies that prioritize the prevention of SDH in all
crash configurations and age groups, but particularly in crashes involving elderly. For this specific age group, SDH prevention should be
prioritized over concussion prevention.
The evolving nature of the methods used to diagnose concussion
injuries needs to be considered in longitudinal descriptive analysis of
traffic injury data. In the current study we used the AIS90/98 version to
consistently analyze data from a 15 year period running from 2001 to
2015. Within this period, the AIS05 version was released in 2006
(Gennarelli and Wodzin, 2006), updated in 2008 and incorporated into
NASS-CDS from year 2010. Notably, the entire section on concussive
injuries was replaced in the AIS05/08 version with a streamlined set of
injury descriptors that reflect updated neurotrauma diagnostic terminology (Gennarelli and Wodzin, 2006). Prior to the incorporation of the
AIS05/08 version into NASS-CDS, crash investigators and physicians
underwent training on the new codes, which likely affected their tendency to code concussions also when using the AIS90/98, particularly
from year 2010. To control for any possible effect this change may have
had on the NASS-CDS data, a variable to control for crash years prior or
after 2010 was incorporated into the odds ratio analysis (Table 4). By
doing so, we confirmed a significant increase by a factor of 1.6 in the
likelihood of a case to be reported as a concussion in case years 2010
and later. Hence, the apparent increase of concussion risk observed in
Fig. 2 can be attributed to the evolving nature of concussion diagnosis,
rather than vehicle or occupant related factors.
Concussions coded based on self-reported or witness-based symptoms may not be reliable, particularly for the mildest forms of concussion that are manifested by symptoms that only last for a few seconds since the patients and the witnesses may miss or exaggerate the
symptoms. In addition, concussions might often be under-reported in
hospital records since discharge records are often coded by the billing
offices and may not capture all instances of concussion, especially when
following the AIS coding rules. The possible effect of diagnosis or
hospital protocols related under or over-reporting of concussions in our
study can be expected to be relatively small since we excluded the
mildest (AIS1) concussions and focused on AIS2-3 concussions which
symptoms tend to last longer and their diagnosis can also be based on
neurological deficits that are objectively documented in the hospital
charts. Nevertheless, some of the AIS90/98 codes we grouped as concussion cannot discriminate between unconsciousness durations of few
seconds from longer periods up to one hour. Hence, although the current study provides a valuable estimation of the proportions of AIS2
concussions based on a consistent data set including 15 years of data
and identifies important gender and age trends in concussion, there is
room for further improvement of the estimations. It is noted that the
AIS05/08 code not only provides a streamlined set of injury descriptors
that will increase the reliability of concussions diagnoses, but also allows to discriminate loss of consciousness of less than 30 min from
durations between 30 min and 1 h. As additional data corresponding
with new years are added to the NASS-CDS data base and sufficient
amount of data coded with the AIS05/08 version becomes available,
more refined studies that better control for the effect of self-reported
symptoms in moderate and serious concussions will become possible.
Female occupants are at higher risks than male occupants of sustaining a concussion in frontal car crashes. To the best knowledge of the
authors, the current study is the first that identifies an increased risk for
concussion in females in MVCs. More precisely, after controlling for
crash severity, crash year 2010, sex and age, belted female occupants
were estimated to be 1.5 times more likely to sustain a concussion than
belted male occupants. It is known from sports medicine that female
athletes may be up to 2.6 times more likely to sustain a concussion than
Table 7
Odds ratio estimates based on logistic regression analysis of SDH cases.
Significant (p < 0.05) point estimates marked with an asterisk. (Crash Year
2001–2015, Car Model Year 2001–2015, Non-ejected, Age 15+, Belted, Side
crashes).
Effect
Estimate
95% Confidence Limits
p-value
delta-V
Sex (Female vs Male)
Elderly (65+ vs 15-64)
Crash Year (2010+ vs 2001/
09)
Model Year (2001/04 vs
2011/15)
Model Year (2005/10 vs
2011/15)
1.09*
0.58
7.95*
0.67
1.07
0.24
3.00
0.24
1.11
1.38
21.00
1.93
< 0.0001
0.1995
0.0004
0.4386
6.65*
1.34
32.85
0.0030
2.70
0.51
14.32
0.9178
Finally, from the 2,389,173 weighted occupants that were involved
in a side crash with known delta-V, 4628 sustained a SDH. Table 7
summarizes the results from applying logistic regression to the occurrence of SDH in this subset of side crashes. Delta-V, occupant age and
model year 2010 + significantly contributed to predict the probability
of SDH, while sex and crash year 2010 did not. After adjusting for the
effects of delta-V, a belted elderly occupant involved in a side crash is
estimated to be almost 8 times more likely to sustain a SDH than a nonelderly belted adult. Further, the risk for an occupant to sustain an
injury in a vehicle produced between 2001 and 2004 was 6.6 times
higher than that of an occupant in a vehicle produced in 2011 or later.
4. Discussion
Frontal and side crashes together comprise the majority of brain
injuries but largely differ between them in terms of brain injury risks.
Compared to frontal crashes, occupants in side crashes are most frequently located close to the intruding structure, which leaves less room
and time for crashworthiness and restraint systems interventions to
protect the occupants. In addition, the introduction and evolution of
safety regulations and ratings for side impacts occurred later than in
frontal impacts. For example, the effectiveness of seatbelts and frontal
airbags in reducing frequencies of skull and AIS2+ brain injuries in
frontal crashes was shown based on analysis of 1991 to 1998 crash data
(Pintar et al., 2000), while the effectiveness of side airbags in side
crashes became available in later studies based on crash data from years
not earlier than 1997 and onwards (Yoganandan et al., 2007; Griffin
et al., 2012; Sunnevång et al., 2015). All this is reflected in the current
study. First, the risk for all brain injury categories, except for contusions, was higher in side crashes than in frontal crashes (Table 3).
Second, vehicle model year was not a significant contributor to explain
neither concussions nor SDHs risks in frontal crashes (Tables 4 and 6),
while it was for both injury categories in side crashes (Tables 5 and 7).
This supports the idea that the most significant safety improvements in
frontal crashes occurred prior to the 2001–2015 period analyzed in this
study, while the most significant side crash safety improvements have
occurred within this period. Despite the specific improvements identified for side crashes, occupants in side crashes are still at higher risks
than in frontal and the improvements are insufficient to produce overall
and significant brain injury risk reductions (Fig. 1).
Elderly occupants are at higher risks than non-elderly occupants of
sustaining AIS3 + brain injuries in both frontal (Table 6) and side
(Table 7) crashes. A number of studies in the past have shown age to be
an important risk factor for AIS3+ head injuries (Ridella et al., 2012).
However, relatively few studies have focused on the effects of age on
specific brain injuries. An example is a study that looked into the effect
of age on both isolated and accompanied bleeding injuries by analyzing
NASS-CDS data from years 1993–2007 (Mallory, 2010; Mallory et al.,
2011). That study identified that elderly were particularly vulnerable to
extra-axial bleeding injuries (including SDH, SAH and EDH) (Mallory,
103
Accident Analysis and Prevention 117 (2018) 98–105
J. Antona-Makoshi et al.
5. Limitations
their male counterpart (Zuckerman et al., 2015). The reasons for this
difference are under debate (Rowson et al., 2016); some studies have
shown that there may be a cultural bias that encourages male athletes
to continue practicing the sports activity despite a concussion, while
female athletes may be more concerned about the effects of an injury on
their future health (Torres et al., 2013). Since the arguments for cultural bias do most likely not stand for automotive concussion victims,
the current study supports the idea that females are indeed more likely
to sustain a concussion than males and that the causes for this difference are related to different biomechanical responses to comparable
exposures.
In general, women are more fragile than men, meaning they are
more susceptible to injury in comparable loading conditions, and less
frail, meaning they are less likely to die from comparable injuries. The
increased fragility in women clearly reflects in traffic accident data
analysis studies in which the ratios and risks sustained by women are
higher than those sustained by men (Bose et al., 2011; Kahane, 2012;
Klier et al., 2016). By body regions, increased AIS2 + injury risks in
women have been identified to be a 22% higher for the head, a 45% for
the neck, a 28% for the chest, a 39% for the abdomen and an 80 and
58% for the lower and upper extremities, respectively (Kahane, 2012).
This sex related differences have received particular attention in Whiplash Associated Disorders (WAD) injuries which are predominantly
suffered by young female occupants involved in rear impact crashes.
The risk for WAD in this kind of impacts has been shown to be between
1.5 and 3 times higher in female occupants compared to males
(Kullgren and Krafft, 2010) and the underlying biomechanics have been
largely investigated (Sato et al., 2010; Carlsson et al., 2012). Among the
reasons that may affect different response to impacts, neck musculature
strength (Vasavada et al., 2008) and reaction times (Siegmund et al.,
2003), cervical fiber and ligament composition (Stemper et al., 2008),
cervical vertebrae anatomical differences (Stemper et al., 2008), spinal
alignment in seated postures (Sato et al., 2016) and the effect of these
differences in simulated crashes (Sato et al., 2017) have been addressed. These studies have motivated the development of biomechanics research tools that better represent female average populations such as anthropomorphic devices (Schick et al., 2010; Carlsson
et al., 2012; Linder et al., 2013) and human body numerical models
(Kitagawa et al., 2015; Östh et al., 2017). The increased risk for concussion in female occupants we found in the current study was statistically significant only in frontal impacts and hence cannot be generalized to other crash configurations. However, it is reasonable to
think that many of the sex differences that have shown to influence
neck injury biomechanics in rear crashes may also play a role to understand increased concussion injury risk in female occupants in different crash configurations. Expanding WAD biomechanics research to
crash configurations other than rear and concussion biomechanics research to all crash configurations may lead to a better understanding of
the craniocervical injury biomechanics (Elkin et al., 2016). It remains
for future work to clarify if these differences in injury risks are simply
related to how restraint systems installed in passenger vehicles perform
for male and female occupants, if the differences in risks are associated
to inherent anatomical previously suggested for WAD injuries, or if the
differences are caused by other differences between sexes not considered in neck injury studies such as different mechanical properties of
the brain tissue (Finan et al., 2017). Regardless of the research strategies adopted, further studies on the effect of sex on the risk of concussion in traffic crashes are necessary. Test methods that assess the risk
of concussion should be developed and implemented. These methods
should account for the fact that most concussions occur at low and
moderate crash severities (Viano and Parenteau, 2015) and should
prioritize injury prediction in females over males particularly in frontal
crashes.
NASS-CDS is a sample of crashes for which the vehicles are towed
away from the accident scene. This implies that lowest severity crashes
and the brain injuries that could potentially occur in those crashes are
not included in the database. To mitigate this limitation, we removed
AIS1 Concussions from the analysis and focused on AIS2-3 concussions,
which are more likely to occur at the moderate crash severities that are
more frequent in NASS-CDS as compared to low severity crashes.
Nevertheless, AIS2-3 concussions may also occur at very low severity
crashes, which would imply that some injuries cases may be missing in
the analyzed data sample. Hence, the estimations of AIS2-3 concussions
presented in the current study may be considered an underestimation of
the real-world scenario.
In this study, frequency of a particular injury was defined as the
number of occupants who sustained that particular injury as their
highest severity brain injury (MAISbrain). Focusing on the maximum
severity injury is common practice for prioritization purposes in the
development of prevention strategies, which this study mainly targets
at. We are confident that this approach did not considerably affect the
results and conclusions related to concussions; there was an accompanying brain injury of higher severity in less than two percent of the
occupants that sustained an AIS2-3 concussion (Table 3 in contrast with
Table A7). However, when TBIs of high injury severity are accompanied
by other injuries it is well known that the occupant’s injury outcome is
affected (Styrke et al., 2007). Hence, incorporating all the brain injuries
that an occupant sustains may suggest slightly different injury prevention strategies as compared to this study and will provide greater
effectiveness in real-world traffic (Mallory, 2010; Mallory et al., 2011).
Such an incorporation call, however, for future research on injury
outcome measures.
Perhaps the most relevant limitation of this study is that it is based
on a single database of vehicle crashes in the US. These database and
crashes do not necessarily reflect the traffic environment, vehicle or
occupant characteristics of other countries, and hence may compromise
the direct applicability of this study to other countries. Future work will
address cross comparison of real-world crash data from international
sources.
6. Conclusion
NASS-CDS data were analyzed to estimate the frequency and risk of
all AIS2+ brain injuries sustained by occupants in car crashes that
occurred in the US from years 2001 to 2015. The data were analyzed
considering crash year, crash type, crash severity, car model year, belt
usage and the victim’s age and sex. It appears that adopted occupant
protection strategies are insufficient to achieve a significant decrease in
risk of both life-threatening brain injuries and concussions injuries in
car occupants. Further effort is needed to develop occupant and injury
specific strategies for the prevention of brain injuries. This study suggests that future prevention strategies need to prioritize life-threatening
vasculature brain injuries, particularly in elderly occupants, and concussion injuries, particularly in female occupants.
Acknowledgements
Funding for this research was provided by Chalmers Area of
Advance in Transport and by the Japan Automobile Research Institute.
David Viano and Chantal Parenteau are acknowledged for providing
guidance and expertise before the initiation of this study. The members
of the Impact Biomechanics Experts Committee of the Society of
Automotive Engineers of Japan (JSAE) are acknowledged for providing
feedback at intermediate steps of this study.
104
Accident Analysis and Prevention 117 (2018) 98–105
J. Antona-Makoshi et al.
References
Ridella, S.A., Rupp, J.D., Poland, K., 2012. Age-related differences in AIS 3+ crash injury
risk, types, causation and mechanisms. IRCOBI Conference Proceedings.
Rowson, S., Bland, M.L., Campolettano, E.T., Press, J.N., Rowson, B., Smith, J.A., Sproule,
D.W., Tyson, A.M., Duma, S.M., 2016. Biomechanical perspectives on concussion in
sport. Sports Med. Arthrosc. 24 (3), 100–107.
Sato, F., Antona, J., Ejima, S., Ono, K., 2010. Influence on cervical vertebral motion of the
interaction between occupant and head Restraint/Seat, based on the reconstruction
of rear-end collision using finite element human model. IRCOBI Conference
Proceedings.
Sato, F., Odani, M., Miyazaki, Y., Nakajima, T., Makoshi, J.A., Yamazaki, K., Ono, K.,
Svensson, M., Östh, J., Morikawa, S., 2016. Investigation of whole spine alignment
patterns in automotive seated posture using upright Open MRI systems. IRCOBI
Conference.
Sato, F., Odani, M., Miyazaki, Y., Yamazaki, K., Östh, J., Svensson, M., 2017. Effects of
whole spine alignment patterns on neck responses in rear end impact. Traffic Inj.
Prev. 18 (2), 199–206.
Schick, S., Kullgren, A., Tomasch, E., Jakobsson, L., Linder, A., Gales, N., Hell, W.,
Schmitt, K., 2010. Basics for developing a female occupant model for investigating
cervical spine distortion injury (CSD). ESAR Conference.
Siegmund, G.P., Sanderson, D.J., Myers, B.S., Inglis, J.T., 2003. Rapid neck muscle
adaptation alters the head kinematics of aware and unaware subjects undergoing
multiple whiplash-like perturbations. J. Biomech. 36 (4), 473–482.
Stemper, B.D., Yoganandan, N., Pintar, F.A., Maiman, D.J., Meyer, M.A., DeRosia, J.,
Shender, B.S., Paskoff, G., 2008. Anatomical gender differences in cervical vertebrae
of size-matched volunteers. Spine 33 (2), E44–E49.
Styrke, J., Stålnacke, B.-M., Sojka, P., Björnstig, U., 2007. Traumatic brain injuries in a
well-defined population: epidemiological aspects and severity. J. Neurotrauma 24
(9), 1425–1436.
Sunnevång, C., Sui, B., Lindkvist, M., Krafft, M., 2015. Census study of real-life near-side
crashes with modern side airbag-equipped vehicles in the United States. Traffic Inj.
Prev. 16 (Suppl. 1), S117–S124.
Takhounts, E.G., Craig, M.J., Moorhouse, K., McFadden, J., Hasija, V., 2013.
Development of brain injury criteria (BrIC). Stapp Car Crash J. 57, 243.
Talmor, D., Thompson, K.M., Legedza, A.T., Nirula, R., 2006. Predicting severe head
injury after light motor vehicle crashes: implications for automatic crash notification
systems. Accid. Anal. Prev. 38 (4), 767–771.
Torres, D.M., Galetta, K.M., Phillips, H.W., Dziemianowicz, E.M.S., Wilson, J.A., Dorman,
E.S., Laudano, E., Galetta, S.L., Balcer, L.J., 2013. Sports-related concussion anonymous survey of a collegiate cohort. Neurol. Clin. Pract. 3 (4), 279–287.
Vasavada, A.N., Danaraj, J., Siegmund, G.P., 2008. Head and neck anthropometry, vertebral geometry and neck strength in height-matched men and women. J. Biomech.
41 (1), 114–121.
Viano, D.C., Parenteau, C.S., 2015. Concussion, diffuse axonal injury and AIS4+ head
injury in motor vehicle crashes. Traffic Inj. Prev. 16 (8), 747–753.
World Health Organization, 2013. WHO Global Status Report on Road Safety 2013:
Supporting a Decade of Action.
Yoganandan, N., Pintar, F.A., Zhang, J., Gennarelli, T.A., 2007. Lateral impact injuries
with side airbag deployments—a descriptive study. Accid. Anal. Prev. 39 (1), 22–27.
Zaloshnja, E., Miller, T., Romano, E., Spicer, R., 2004. Crash costs by body part injured,
fracture involvement, and threat-to-life severity, United States, 2000. Accid. Anal.
Prev. 36 (3), 415–427.
Zhang, F., Chen, C.-L., 2013. NASS-CDS: Sample Design and Weights.
Zuckerman, S.L., Kerr, Z.Y., Yengo-Kahn, A., Wasserman, E., Covassin, T., Solomon, G.S.,
2015. Epidemiology of sports-related concussion in NCAA athletes from 2009-2010
to 2013-2014 incidence, recurrence, and mechanisms. Am. J. Sports Med. 43 (11),
2654–2662.
Bener, A., Omar, A.O.K., Ahmad, A.E., Al-Mulla, F.H., Abdul Rahman, Y.S., 2010. The
pattern of traumatic brain injuries: a country undergoing rapid development. Brain
Inj. 24 (2), 74–80.
Bose, D., Segui-Gomez, M., Crandall, J.R., 2011. Vulnerability of female drivers involved
in motor vehicle crashes: an analysis of US population at risk. Am. J. Public Health
101 (12), 2368–2373.
Bruns, J., Hauser, W.A., 2003. The epidemiology of traumatic brain injury: a review.
Epilepsia 44 (s10), 2–10.
Carlsson, A., Siegmund, G.P., Linder, A., Svensson, M.Y., 2012. Motion of the head and
neck of female and male volunteers in rear impact car-to-car impacts. Traffic Inj.
Prev. 13 (4), 378–387.
Carroll, L.J., Cassidy, J.D., Cancelliere, C., Côté, P., Hincapié, C.A., Kristman, V.L., Holm,
L.W., Borg, J., Nygren-de Boussard, C., Hartvigsen, J., 2014. Systematic review of the
Prognosis after Mild traumatic brain injury in adults: cognitive, psychiatric, and
mortality outcomes: results of the International Collaboration on Mild TBI Prognosis.
Arch. Phys. Med. Rehabil. 95 (3), S152–S173.
Eigen, A.M., Martin, P.G., 2005. Identification of real world injury patterns in aid of
dummy development. 19th ESV Conference.
Elkin, B.S., Elliott, J.M., Siegmund, G.P., 2016. Whiplash injury or concussion? A possible
biomechanical explanation for concussion symptoms in some individuals following a
rear-end collision. J. Orthop. Sport Phys. Ther. 46 (10), 874–885.
Finan, J.D., Sundaresh, S.N., Elkin, B.S., McKhann, G.M., Morrison, B., 2017. Regional
mechanical properties of human brain tissue for computational models of traumatic
brain injury. Acta Biomater. 55, 333–339.
Gennarelli, T.A., Wodzin, E., 2006. AIS 2005: a contemporary injury scale. Injury 37 (12),
1083–1091.
Griffin, R., Huisingh, C., McGwin Jr., G., Reiff, D., 2012. Association between side-impact
airbag deployment and risk of injury: a matched cohort study using the CIREN and
the NASS-CDS. J. Trauma Acute Care Surg. 73 (4), 914–918.
Kahane, C.J., 2012. Relationships between Fatality Risk, Mass, and Footprint in Model
Year 2000-2007 Passenger Cars and LTVs ? Final Report. NHTSA Technical Report
DOT HS 811 665.
Kitagawa, Y., Yamada, K., Motojima, H., Yasuki, T., 2015. Consideration on gender difference of whiplash associated disorder in low speed rear impact. IRCOBI Conference
Proceedings.
Klier, W., Freienstein, H., D’Addetta, G.H., Köhler, A., Reckziegel, B., Shiozawa, K.,
Schulz, A., Cuvillier, M., 2016. Interior sensing for occupant protection. In: Airbag
2016, 13th International Symposium on Sophisticated Car Safety. Mannheim,
Germany.
Kullgren, A., Krafft, M., 2010. Gender analysis on whiplash seat effectiveness: results from
real-world crashes. IRCOBI Conference Proceedings.
Linder, A., Schick, S., Hell, W., Svensson, M., Carlsson, A., Lemmen, P., Schmitt, K.-U.,
Gutsche, A., Tomasch, E., 2013. ADSEAT–Adaptive seat to reduce neck injuries for
female and male occupants. Accid. Anal. Prev. 60, 334–343.
Mallory, A., 2010. Head injury and aging: the importance of bleeding injuries. Ann. Adv.
Automot. Med. 54 (5).
Mallory, A., Herriott, R., Rhule, H., 2011. Subdural hematoma and aging: crash characteristics and associated injuries. 22nd ESV Conference.
Östh, J., Mendoza- Vazquez, M., Linder, A., Svensson, M.Y., Brolin, K., 2017. The VIVA
OpenHBM finite element 50th percentile female occupant model: whole body model
development and kinematic validation. IRCOBI Conference Proceedings.
Pickrell, T.M., 2017. Seat Belt Use in 2016 Use Rates in the States and Territories. NHTSA
Technical Report DOT HS 812 417.
Pintar, F., Yoganandan, N., Gennarelli, T., 2000. Airbag effectiveness on brain trauma in
frontal crashes. AAAM Conference.
105